Skip to main content

Full text of "computersAndAutomation :: 195910"

See other formats


OCTOBER 

1959 

• 

VOL 8 - NO. 10 



Conversation With a Computer 
Important Applications of Computers 
Machine Translation From English to Russian 
Air Traffic Control in the Year 1980 





COMPUTER ENGINEERS 

Bendix Aviation Corporation's Eclipse-Pioneer Division 

Invites You 

to Participate in a Newly-Bxp ancle d Program in 

TRANSISTORIZED AIRBORNE 
DIGITAL COMPUTING SYSTEMS 



The expansion of our Advanced Research and Development program 
in solid state airborne digital computation and control has created a 
number of outstanding opportunities for qualified computer engineers. 
If you are interested in working on new digital concepts with a leader 
in the airborne computing and control system fields, and if you are cur- 
rently qualified in one of the following areas . . . 



Digital Systems Synthesis 
Logical Design 
High Speed Switching 
Logical Implementations 
Magnetic Memory 



Input-Output Devices 
Transistorized Circuitry 
Pulse Techniques 
Miniaturization 
Micro-Miniaturization 



... we suggest that you send a brief resume of your experience and educa- 
tional background to Dr. Arthur S. Robinson, or telephone Engineering 
Personnel at ATlas 8-2000 in New Jersey, or BRyant 9-8541 in New York. 

ECLIPSE-PIONEER DIVISION 

AVIATION CORPORATION 

Route 46 at 17, Teterboro, New Jersey 
(Located in suburban New Jersey, 10 minutes from New York City) 



COMPUTERS and AUTOMATION for October, 1959 





Sound decisions depend upon fast, accurate facts 

Management decisions must be based upon up-to-the-minute facts. With so much at stake, there can be no doubt 
of accuracy, no loss of time, in reducing mountains of data to decisionable" facts. The Philco Transac S-2000 all- 
transistor data processing system provides unequalled accuracy and speed. First in fully transistorized data p 9 

Transac S-2000 offers outstanding economies in installation, operation and maintenance. It assures freedom from 
obsolescence . . . being asynchronous it. can be updated without changing existing programs. Before making your 
decision on a large scale data processing system, get the facts on Transac ... the computer that changed the industry. 



VmnAML 



s-2000 




GOVERNMENT AND INDUSTRIAL. DIVISION • 



4 700 



W ISSAH 1CKON AVENUE. 



PHILADELPHIA A A . 



M P U T E R 

AUTOMATION 



s 



ROBOTS 



c o 

and 

DATA PROCESSING 



CYBERNETICS • 



Volume 8 
Number 10 



OCTOBER, 1959 



Established 
September 1951 



1 


’Edmund C. Berkeley" Editor 

H. Jefferson Mills, Jr. Assistant Editor: 
Neil D, Macdonald Assistant Editor 

SERVICE and SALES DIRECTOR 

Milton L. Kaye MUrray Hill 2-4194 

535 Fifth Ave. New York 17, NY.' 

CONTRIBUTING EDITORS 
Andrew D. Booth 

Ned Chapin 

John W. Carr, III , * , , 

Alston S. Householder 

ADVISORY COMMITTEE 
Morton M. Astra a an 

Howard T. Engstrom; 
George E. Forsythe 

Richard W. Hamming 
Alston S. Householder 

Herbert F. Mitchell, Jr. 

Samuel B. Williams 

- 

ADVERTISING REPRESENTATIVES 
Middle Atlantic States Milton L. Kaye 
535 Fifth Ave. New York , 17, N.Y. 

MUrray Hill 2-4194 

Washington, 6, D.C. Robert Cadel 

; 1519 Connecticut Ave. ' Columbia 5-9727 
j San Francisco J> A. S. Babcock 

j 605 Market St. • YUkon 2-3954 

\ Los Angeles 5 W.' F. Green 

j 439 S. Western Ave. DUnkirk 7-8135 

| Hi >e where , The Publishes 

f • Berkeley Enterprises, Inc. 

i 815 Washington St., Newtonville 60, Mass 
\ . DEcatur 2-5453 or 2-3928 

L_: ...... „ ' • 





COMPUTERS and AUTOMATION is published monthly 
at 815 Washington St., Newtonville 60, Mass., by Berkeley 
Enterprises, Inc. Printed in U.S.A. 

SUBSCRIPTION RATES: (United States) $5.50 for 1 

year, $10.50 for 2 years; (Canada) $6.00 for 1 year, $11.50 
for 2 years; (Foreign) $6.50 for 1 year, $12.50 for 2 years. 

Address all Editorial and Subscription Mail to Berkeley 
Enterprises, Inc., 815 Washington St., Newtonville 60, 
Mass. 

ENTERED AS SECOND CLASS MATTER at the Post 
Office at Boston, Mass. 

POSTMASTER: Please send all Forms 3579 to Berkeley 
Enterprises, Inc., 815 Washington St., Newtonville 60, 
Mass. 

Copyright © 1959, by Berkeley Enterprises, Inc. 
CHANGE OF ADDRESS: If your address changes, please 
send us both your new address and your old address (as 
it appears on the magazine address imprint), and allow 
three weeks for the change to be made. 



SOME IMPORTANT APPLICATIONS 
OF COMPUTERS 



Braille Translation 12 

Sixty Seconds from Engine Testing to Final Evaluation 13 

More Adequate Selection of Candidates for Jobs . . 14 

All Daily Work for Bank Check Accounting . . . 14 

Sante Fe Railway Reservation System . . . . 14 

Casualty Insurance Actuarial Calculations . ... 16 

Ray Tracing in a Cassegrainian Optical System . . 1 6 

Automatic Machining of Parts with Two- and Three- 

Dimensional Contours 1 6 

Predicting the Soundness of Foundry Castings . . 17 

Computers Simulating Men 17 

Flood Control by Computer 18 

Automatically Shaping Noncircular Gears ... 18 

Reservoir Filling 18 

Sea Rescue Operations 18 

Weather Forecasting 18 

FRONT COVER 

First Successful Operation of a Practical Magnetic Film 
Memory 1, 6 

ARTICLES 

Conversation with a Computer, L. E. S. Green, E. C. 
Berkeley, and C. C. Gotlieb 9 



Machine Translation Methods, and Their Application to 
Translation from English to Russian, I. K. Belskaya 20 
Air Traffic Control in the Year 1980, Vernon I. Weihe 32 

READERS’ AND EDITOR’S FORUM 

Correction of Title — "A General Problem-Solving Pro- 
gram for a Computer 

Working Group on Socially Desirable Applications of 

Computers 

On Freedom to Publish, James A. Cook, etc. . 

I960 Western Joint Computer Conference — Invitation 

for Papers 

Notice from a Committee on Data Systems Languages, 

A. E. Smith 



REFERENCE INFORMATION 

Calendar of Coming Events 11 

Early Issues of "Computers and Automation” . . .31 

Books and Other Publications, M. M. Berlin ... 39 

Survey of Recent Articles, M. M. Berlin .... 45 

INDEX OF NOTICES 

Advertising Index 46 

Back Copies 39 

Bulk Subscriptions 39 

The Computer Directory and Buyers’ Guide, 1959 . 

see June, page 91 

Manuscripts see August, page 28 

Reference and Survey Information . see August, page 29 

Who’s Who Entry Form . . . see August, page 32 



COMPUTERS and AUTOMATION for October, 1959 



6 

6 

30 

30 

31 



4 







NOW FROM AM REX; 



The world’s most advanced scientific data recorder. More useful at a countdown than at a concerto, this 
third-generation machine — like its cousins and ancestors — brings you repeat performances at a fraction of 
the cost of re-staging the original, perhaps unique event — albeit satellite launching instead of symphony. 
Write Ampex Instrumentation, 934 Charter St., Redwood City, Calif., for a brochure on the distinguished 
new FR-600. It’s solid-state . . . extremely reliable . . . extremely precise. You expect it from / j r , F*f 






and 

Editor's Forum 



Readers' 



FRONT COVER: FIRST SUCCESSFUL OPERATION 
OF A PRACTICAL MAGNETIC FILM MEMORY 

John A. Kessler 

Mass. Inst, of Techn., Lincoln Lab. 

Lexington 73, Mass. 

The front cover shows a high-speed magnetic film 
memory now in operation as a part of the TX-2 digital 
computer at the MIT Lincoln Laboratory. It is be- 
lieved to be the first in successful practical operation. 
Its performance has been entirely satisfactory since its 
installation in July 1959- It has a capacity of 32 ten-bit 
words, suitable for evaluation testing, and serves as an 
experimental prototype for larger units. This new mem- 
ory, and the TX-2 computer of which it is a part, were 
developed by Lincoln Laboratory under Air Force con- 
tract, with the joint support of the Army, Navy, and 
Air Force. 

The read-and-write cycle time of 0.8 microseconds is 
consistent with the speed of the computer itself, al- 
though bench tests demonstrated successful operation at 
a cycle time as short as 0.4 microseconds. Net driving 
current for writing is 150 milliamperes, and one-milli- 
volt output signals are obtained from individual mem- 
ory elements. 

Each memory element is a circular spot of Permalloy 
film (82 percent nickel, 18 percent iron) 750 Angstroms 
thick, 1.6 millimeters in diameter, centered 2.5 milli- 
meters apart. The spots are deposited by evaporation on 
a flat glass substrate, 0.1 millimeter thick, in 16 x 16 
unit arrays. The complete memory unit as installed in 
TX-2 and one of the experimental arrays are shown in 
the two figures. The transistor drive and sense circuits 
can be seen surrounding the memory. 




This is a magnetic film memory array of the type installed in 
the TX-2 computer at MIT Lincoln Laboratory. It is experi- 
mental. The memory elements are circular spots of film of 
Permalloy (82 percent nickel, 18 percent iron). The thickness 
is 750 Angstroms. It was deposited on a thin glass plate. Each 
memory element is 1.6 millimeters in diameter; the center-to- 
center spacing is 2.5 millimeters. 



A thin film memory has several potential advantages 
over the familiar ferrite toroidal core memory: faster 
cycle time, lower power dissipation, greater compact- 
ness, and simpler fabrication. The unit now in opera- 
tion confirms these expectations, although none of these 
factors has been fully exploited in this first develop- 
mental model. 

CORRECTION OF TITLE — "A GENERAL 
PROBLEM-SOLVING PROGRAM 
FOR A COMPUTER” 

In the July, 1959, issue of Computers and Automa- 
tion, a paper "A Report on a General Problem-Solving 
Program for a Computer” by A. Newell, J. C. Shaw, 
and H. A. Simon was printed starting page 10 with 
the title omitting the words "A Report on,” and thus 
implying, according to the authors, more finality to the 
report than the authors desire. 

The title of this paper should be corrected to "A 
Report on a General Problem-Solving Program for a 
Computer.” 

WORKING GROUP ON SOCIALLY DESIRABLE 
APPLICATIONS OF COMPUTERS 

At the Association for Computing Machinery Meeting 
in Cambridge, Mass., on Sept. 1, 1959, an afternoon 
session was held on "Social Aspects of Computing.” 
After one paper was presented ("The Computing Ma- 
chine — Slave Labor in a Free Society,” by H. M. El- 
liott), a Panel Discussion on the Social Responsibilities 
of Computer People was begun by Arvid W. Jacobson, 

H. R. J. Grosch, and Louis Sutro, and continued by a 
great many members of the audience. 

The discussion lasted vigorously for nearly 3 hours. 

During the session, suggestions were made by Dr. 
Herbert R. J. Grosch, consultant, Evan Herbert, editor 
of Data Control, and E. C. Berkeley that we organize 
a "working group on socially desirable applications of 
computers,” consisting of those persons who would sign 
up at the front of the room at the end of the session, and 
any other interested computer people. Twenty-nine 
persons signed up at the time as interested in working 
in this group, and the following questionnaire has been 
sent out to them. 

Any interested reader of Computers and Automation 
is invited to complete this questionnaire, and send it in 
also. 

The ACM Committee on the Social Responsibilities of 
Computer People of course still continues to exist and 
function, but with a difference in emphasis. 

WGSDAC Questionnaire 

(may be copied on any piece of paper; and any interested 
computer person that you know may fill in a copy also) 

I. What do you suggest as the main purposes for this 

group ? 

[Please turn to page 30] 

COMPUTERS and AUTOMATION for October, 1959 



6 




A Hughes representative will gladly demonstrate the 
MEMO-SCOPE. Address requests to: hughes products. 
Industrial Systems Division, International Airport 
Station, Los Angeles 45, Calif. For export information, 
write: hughes international. Culver City, Calif. 



Attempting to analyze elusive non-recurring tran- 
sients has sent the best of engineers “to the showers.” 
Conventional practices for monitoring or examining 
spurious signals or noise factors do not permit accu- 
rate investigations. The result is a wasteful expense 
of your time, your nerves, and your company’s money. 

SOLUTION: The Hughes memo-scope® oscilloscope 
freezes wave forms until intentionally erased. Selected 
transient information may be triggered externally or 
internally and retained for viewing. Successive wave 
forms may be written above, below or directly upon 
the original information. 



HUGHES MEMO-SCOPE OSCILLOSCOPE 

SWEEP SPEED FOR STORAGE: 

10 microseconds to 10 seconds per division (0.3"). 

FREQUENCY RESPONSE: DC to 250 KC down 3 db. 

SENSITIVITY: 

10 millivolts to 50 volts per division or with optional high sensi- 
tivity preamplifier 1 millivolt to 50 volts per division. 

APPLICATIONS: 

Trouble shooting data reduction equipment... switch and relay 
contact study. .. ballistics and explosives research ... ultrasonic 
flaw detection ... physical testing — shock — stress — strain. 



INDUSTRIAL SYSTEMS DIVISION 

I 

I 

Creating a new world with ELECTRONICS 

HUGHES PRODUCTS 



) 19 59, HUGHES AIRCRAFT COMPANY 



SEMICONDUCTOR DEVICES • STORAGE TUBES AND DEVICES • MICROWAVE TUBES • VACUUM TUBES AND COMPONENTS • CRYSTAL FILTERS • MEMO-SCOPE® OSCILLOSCOPES • INDUSTRIAL CONTROL SYSTEMS 



COMPUTERS and AUTOMATION for October, 1959 



7 





Join the leader of the new generation of computers ! 



METHODS ANALYSTS 
FOR HONEYWELL 800 



Make your big move now. All- 
transistorized Honeywell 800 is rapidly 
winning customers coast to coast. Make a 
name for yourself now, working with this 
most advanced data processing system in 
the medium to large-scale field. 

Enjoy management's full sup- 
port. Honeywell is dedicated to continuing 
progress in computer design and service. 
All data processing efforts are solidly sup- 
ported right down the line. 

Advance in your own field. 

Honeywell offers you a chance for high 
advancement in your chosen field. No need 
to transfer to another department to get 
ahead. 



Can you qualify? 

★ BS or MS in any science, math, statistics, 
accounting or business 

★ 3 to 5 years computer and applications 
experience 

★ Willingness to relocate, if necessary 

Complete training on Honeywell 800 in 
Boston precedes assignment to locations in 
New York, Boston, Washington, Chicago, 
Detroit, Los Angeles and other major cities. 

Write in confidence and send resume to: 
John B. Koenig 

Professional Employment Supervisor 
Minneapolis-Honeywell, Datamatic Div. 
Newton Highlands 61, Mass. 



8 ways to identify the leader of the new generation of computers 



1 


2 


3 


4 


5 


6 


7 


8 


EFFICIENCY 


ECONOMY 


EXPANDABILITY 


RELIABILITY 


VERSATILITY 


ADAPTABILITY 


SERVICE 


BACKING 


You can do up 


You can 


You can 


You can cut 


You can do 


You 


You can expect 


You can 


to 8 different, 


process more 


increase 


costs 


both 


can simplify 


complete 


count on the 


independently 


data per dollar 


Honeywell 800 


of maintaining 


business data 


problem 


training, 


continuing 


programmed 


in a working 


capacity 


accuracy with 


processing and 


preparation 


programming 


support of a 


jobs — 


day, on small 


indefinitely 


a new kind of 


scientific 


with 


assistance, 


company 


simultaneously 


volume 


in small, 


automatic 


computation — 


Honeywell’s 


maintenance 


dedicated to 




applications 


economical 


error correction 


simultaneously 


automatic 


and 


the advance of 




or large 


stages 






programming 
aids and 


applications 


transistorized 








f f - 


" .. 


know-how 


data processing 


5 


HI 




; f ” 




library of 
routines 


from 

Honeywell's 
data processing 
specialists 


lW mm « 










1 




- HMMF 





8 



COMPUTERS and AUTOMATION for October, 1959 






CONVERSATION WITH A 

COMPUTER 



L. E. S. Green, KCS Limited, Toronto, Can. 

E. C. Berkeley, Berkeley Enterprises, Inc., 
Newtonville 60, Mass, 
and 

C. C. Gotlieb, Computation Centre, Univ. of Toronto, 
Toronto, Can. 



Introduction 

Can a computer carry on a conversation? 

About 1938 the English mathematician Turing sug- 
gested that a machine could be said to be capable of 
thinking if it could carry on a conversation with a human 
being in another room, in such a way that the human 
being could not tell if he were conversing with a 
machine or with another human being. Since that time, 
this definition of a computer’s thinking has stood as 
a challenge to computer people. 

To make a beginning on the problem, it is clearly 
desirable to restrict it to some relatively simple and 
common subject such as the weather. About 1954 John 
W. Carr, III, now director of the computing center at 
the University of North Carolina in a course on com- 
puters for students at the University of Michigan, as- 
signed as an optional problem for additional credit the 
problem of programming a machine to carry on a con- 
versation about the weather. 1 But no student took up 
the challenge. 

In December, 1958, the problem (as phrased and an- 
alyzed by E. C. Berkeley as a part of a study on language, 
ideas, and computers) was discussed with the Compu- 
tation Centre of the University of Toronto. Dr. C. C. 
Gotlieb of the Centre classified it as a frontier problem 
having interest ; and it was placed in the hands of 
L. E. S. Green, mathematician and programmer, to an- 
alyze, program, and code. 

The first preliminary program ran in April, 1959, and 
showed a surprising degree of success, but also some 
"stupid” or "deaf” replies. The second preliminary 
program ran in August, 1959, and was a good deal more 
satisfactory; but at present writing, still contains some 
undesirable responses for such reasons as failure to 

1 Excerpt from Problems For Students of Computers, by John 
W. Carr, III, published in the February 1955 issue of 
Computers and Automation (Vol. 4, No. 2, pp. 6-8): 

21. Limited Conversation 

Devise a program that will make M I D A C perform what 
appears to be a reasonably lucid conversation about one particular 
subject (for example, the weather). 

[So far no student has tried problem 21, which is a difficult 
problem indeed (an obvious challenge to approximate A. M. 
Turing’s definition of a machine’s "thinking”), but that problem 
has caused numerous comments and discussion among members 
of the class. Several tentative schemes have already been pro- 
posed.] 



recognize certain English idioms, and incompleteness 
of syntactical analysis. 

This article is based on a paper by L. E. S. Green 
and E. C. Berkeley presented at the meeting of the Asso- 
ciation for Computing Machinery at Mass. Inst, of 
Technology, Cambridge, on Sept. 2, 1959. 

At the University of Toronto a preliminary program 
for a computer to converse about the weather has been 
written for the IBM Tape 650 in such fashion that a 
large proportion of the remarks made by the computer 
are sensible. 

Assumptions 

In this program, a number of assumptions have been 
made about the nature of the problem of conversation. 
These are as follows: 

(1) Conversations consist of groups of words. 

(2) In any conversation, a reply to a remark is gen- 
erated by the meaning the remark has for the person 
spoken to or listener, the environment in which the re- 
mark is made, the knowledge and experience of the 
listener, or by the interactions among these three 
factors. 

(3) In conversations about the weather, remarks are 
mainly stereotypes, and are made automatically with 
little cogitation on the part of the speaker. The com- 
ment "Some weather we’re having, eh?” is such an 
example. 

(4) A word, or combination thereof, has a meaning 
to a listener, only if it produces an association with an 
experience in the history of the listener. As a corollary 
then, words which give rise to no associations for the 
listener are meaningless to him. 

(5) It is convenient here to classify words into three 
types, which we shall call ordinary words, time words, 
and operator words. Ordinary words are those which 
have meanings even when isolated from a context, such 
as "snow.” Time words describe times of the year, such 
as "October,” or a time relation to the present, such 
as "last week.” Operator words, here, have no meaning 
when separated, but when taken in their context they 
alter the meaning of other words. Thus, they have a 
function rather than a meaning. For example the word 
"not” by itself is meaningless, but in the statement “The 
sun is not shining,” its function is to change the mean- 
ing of its neighbours. 



COMPUTERS and AUTOMATION for October, 1959 



9 




Inasmuch as these assumptions do not completely 
agree with the properties of conversations, the program 
at the present stage will occasionally generate replies 
which do not make sense. 

Representation of Meaning of Words 
In describing how these assumptions are used in the 
logic of the program I would like to discuss first the 
representation of meaning in the computer program. 
The meaning of an ordinary word is represented by a 
number pair (d, q). 

The second number of the pair, q names the quality 
implicit in the meaning of the word, and the first num- 
ber d, gives a quantitative description of the quality q; 
that is, d is the degree of the quality q. For example 
words describing wet weather are represented as: 



dew 


(1, 


1) 


drizzle 


(3, 


1) 


rain 


(6, 


1) 


cats and doge 


(A 


1) 


downpour 


(9, 


1) 



A meaningless word is represented by the pair (0, 0). 

The function of an operator word is also symbolized 
by a two-number pair, (d, f). Here f designates the 
function to be carried out, and d the degree to which 
the function f must be executed. Examples of operators 
which negate words preceding them in a context are: 
change (1, 13) 

abate (2, 13) 

stop (3, 13) 

Similarly the sense of a time word is represented by 

the pair (d, t), where t gives the type of time, calendar 
or relative, and d again gives the degree. 

For example: 



yesterday 


(1. 10) 


to-day 


(2, 10) 


tomorrow 


(3. 10) 


December 


(1, 9) 


January 


(2. 9) 


February 


(3, 9) 



About 300 English words and their numerical repre- 
sentations are stored in the computer’s memory during 
the program. This set of words constitutes the com- 
puter’s recognition vocabulary — the computer finds these 
and only these words significant in its discourse with 
an observer. Hence we can say that a given word pro- 
duces an association with the computer, if the computer 
is able to find this word in its recognition vocabulary. 

Representation of Meaning of Remarks 

Inasmuch as a remark consists of a group of words, 
a remark may have a meaning also. This meaning is 
determined by carrying out the functions of the operator 
words on the appropriate ordinary words in the remark. 
When this is completed all operators in the remark are 
replaced by the number pair (0, 0) which represents a 
meaningless word. Thus, the original remark, now con- 
sisting of ordinary and time words, is represented by a 
set of non-zero number pairs, and a set of zero number 
pairs. It is the set of non-zero number pairs that we 
call the meaning of the remark. 

Generation of Replies 

The second assumption was that a reply is generated 
by the interaction among the three factors, meaning, en- 
vironment, and machine’s experience. 

The second factor is the environment in which the 

10 



remark is made. This environment for the machine con- 
sists entirely of information about weather — the weather 
for the various months and seasons, and the weather for 
yesterday, to-day, and tomorrow. The latter type is 
told to the machine at the beginning of a run by a 
command to "remember such and such,’’ whereas the 
former is stored internally, since the weather associated 
with the seasons is the same from year to year. In both 
cases, the weather information is stored as a set of 
number pairs. 

The third factor, the machine’s experience, is slightly 
more varied. In addition to the vocabulary, this ex- 
perience consists of a set of weather parameters, such 
as sunny skies and warm breezes, which the machine 
likes, and another set which it dislikes, as determined 
by programmed statements to the machine. Since the 
ordinary words in its vocabulary describe only weather 
and emotion, the machine has a rather narrow outlook. 
Hence, the only aspects of the original remark and of 
the environment which the computer can recognize are 
phases of weather, which it usually associates with a 
time and a preference. 

The third assumption was that conversation about 
the weather is stereotyped. Because of this the com- 
puter is able to reply by selecting a suitable reply frame 
from its memory, filling in a few ordinary weather 
words and time words generated by the interactions just 
considered and punching out the result. By making a 
quantitative evaluation of the degree of interaction the 
machine is able to select the particular reply frame most 
appropriate to the whole situation. In all, there are 
about 350 reply frames, each of which contains a num- 
ber of blanks to be filled by the program when that 
particular frame is chosen as a reply. 

An Example 

Now let us work through a typical example. Let us 
suppose the remark is made to the computer 

"I do not enjoy rain during July.” 

The computer substitutes the numerical representation 
for each of these words as the remark is read in. The 
resulting sentence then appears inside the computer as: 

(0, 0) (0, 0) (3, 14) (3, 11) (6, 1) (0, 0) (8, 9) 
Note that the word not is an operator word with the rep- 
resentation (3, 14). In this remark the word enjoy, (3, 
11), is the operand. The function of (3, 14) is carried 
out on (3, 11) making it into the new word (3, 12), 
dislike, and changing the operator’s word’s representation 
to (0, 0): 

(3, 14) (3,11) (0,0) (3, 12) 

not enjoy — dislike 

The meaning of the remark is now said to be the set of 
non-zero pairs: 

(3, 12) (6, 1) (8, 9). 

This combination of qualities in the original remark 
causes the program to compute the degree of interaction 
between the weather mentioned in the statement: 
rain ( 6, 1 ) 

and the weather associated with the month of July: 
heat ( 7, 5 ) 

blue skies (6, 3) 

It finds that the two sets of weather parameters are in: 
essential disagreement. 

Now the program must choose a reply frame. It notes 
COMPUTERS and AUTOMATION for October, 1959 




that the original remark is characterized by: 
weather (rain) 

emotion ( dislike ) 

calendar time (July) 

statement (period, not question-mark) 
and that the interaction gave a result of: 
disagreement 

In the computer’s memory there is a reply frame corre- 
sponding to this particular combination of qualities and 
interactions. The program then selects this frame, which 
happens to be this: 

"Well, we don’t usually have weather in 

, so you will probably not be disap- 
pointed.” 

This contains two blanks. The program then proceeds to 
fill in the blanks with words originating from the remark, 
environment, or machine’s preferences. In this case the 
reply would then be: 

"Well, we don’t usually have RAINY weather in 
JULY, so you will probably not be disappointed.” 

Some More Examples 

First, we instruct the machine with its current prefer- 
ences and current experience by the statement: 

Remember this month is August. Today we had 
rain and severe heat. Yesterday it was sunny and 
hot. Tomorrow will be cold and cloudy. That 
is all. 

Then for various remarks by the observer O, the machine 
M replies as follows: 

O: Did you enjoy the rain we had this afternoon? 

M: Not really, what I like is sunny weather. 

O: In hot weather one needs at least one bath every 
day. 

M: Yes, I was just out and it was sweltering. 

O: When Christmas comes we will have some colder 
weather. 

M: Cold? Yes, it’s usually pretty frosty during De- 
cember. 

O: Today is clear weather. Will it last? 

M: Let me have no lying. It becomes none but trades- 
men. Now how can overcast weather and the 
clear day you mention coexist? 

O: Do you think atomic weapons testing affects the 
weather ? 

M: Your query is an interrogation of no consequence to 
me. However, how do you feel about today’s 
rain? 

An example of an error caused by the present 
model is: 

O: Can you tell me what the weather is like? 

M: Well I like sunny weather, not the rain we are 
having today. 

Here the computer has incorrectly interpreted the word 
"like” to mean "prefer,” and so the reply does not ade- 
quately correspond to the real meaning of the original 
remark. 

Statistics on the Program 

Some stabiles on the second preliminary program, 
as it is at present, are as follows: 

1. Number of instructions in the program: about 
800 

2. Number of decisions in the program: about 200 

3. Number of words in the machine’s recognition 
vocabulary: 319 but can be easily increased 

COMPUTERS and AUTOMATION for October, 1959 



4. Number of reply frames from which the machine 
can choose the form of an answer: about 350, 
but can be easily increased 

5. Number of words (referring to weather, place, 
and time) that the machine uses to insert into 
blanks in the reply frames: about 75 but can be 
easily increased 

6. Number of parameters of meaning about which 
the machine can react at least partially (such as 
temperature, humidity, precipitation, wind, month 
of year, relative time, place, emotional preference, 
etc.): 17 

7. Time required to generate a reply: about 20 

seconds 

Limitations of the Present Program 
The mathematical model on which the present pro- 
gram is based contains only very limited recognition of 
many parts of English syntax. For example, there is no 
recognition of subject and predicate; nor does it dis- 
tinguish between nouns, adjectives, and verbs. In spite 
of this crudity, the results in passable conversation are 
surprisingly good, which in a way is a commentary on 
the shallowness of ordinary conversation about the 
weather. An unsettled problem is how little additional 
syntax one would have to include so that the program 
would plausibly imitate a man. 

CALENDAR OF COMING EVENTS 

Oct. 5-7, 1959: 5th National Communications Sympo- 
sium, Hotel Utica, N.Y. 

Oct. 12-14, 1959: International Systems Meeting, Royal 
York Hotel, Toronto. 

Oct. 12-14, 1959: 15th National Electronics Conference, 
Hotel Sherman, Chicago, 111. 

Oct. 28-29, 1959: 6th Annual Computer Applications 
Symposium, Armour Research Foundation, Chicago, 
111 . 

Oct. 28-30, 1959: 2nd Annual Equipment Systems Con- 
ference and Exhibit, Conference Building, San Diego, 
Calif. 

Nov. 4-6, 1959: National Automatic Control Conference, 
Sheraton Hotel, Dallas, Texas. 

Nov. 11-13, 1959: 16th National Meeting, Operations 
Research Society of America, Huntington Sheraton 
Hotel, Pasadena, Calif. 

Dec. 1-2, 1959: 4th Midwest Symposium on Circuit 
Theory, Marquette Univ., Milwaukee, Wise. 

Dec. 1-3, 1959: Eastern Joint Computer Conference, 
Statler Hotel, Boston, Mass. 

March 21-24, I960: IRE National Convention, Coli- 
seum and Waldorf Astoria Hotel, New York, N.Y. 
May 2-6, I960: Western Joint Computer Conference, 
San Francisco, Calif. 




11 




SOME 

IMPORTANT APPLICATIONS 

OF 

COMPUTERS 



BRAILLE TRANSLATION 

Translation of printed text into braille at electronic 
speeds is being accomplished with an International Busi- 
ness Machines Corp. electronic computer. 

The procedure creates in minutes an embossed braille 
printing plate suitable for production of books for the 
nation’s 350,000 sightless persons. It was developed by 
IBM mathematicians working in conjunction with the' 
American Printing House for the Blind, leading braille 
book publisher. 



The English-text-into-braille program will compensate 
for the serious shortage of qualified braille translators, 
whose training normally takes two years. A knowledge 
of braille by computer personnel is unnecessary. 

The new method of creating braille was developed 
for a standard IBM 704 data processing system, a large- 
scale computer which has been installed in many indus- 
tries to handle complex scientific and engineering prob- 
lems. The 704 can translate a 300 page book into Braille 
in an hour — a job which would take more than six 
days for a skilled translator. 









































































• 
















































































* 


TH 




*T 


E 


•s 


*T 




H 


A 


*S 




B 


E 


EN 




*P 


*R 


*0 


D 


*5 


C 


e6 




BY 


A 


*N 






















s 








• • 




f • 






• • 






• • 










• • 




























• 








• 


• 






* 


• • 


• • 


* 






* 




• • 


• • 




• 




• 


• 




• 9 








• 


• t 


• 




















I 


R 


M 






7 


0 


A 




CO 


‘P 


’u 


*T 


ER 




AS 






PA 




OF 


A 




* 


B 


*R 


’l 












• 












































T 
































,1 




* 


• • 


4 # 


* 




*• 




/ 




• 




•• 




* 4 


\ 


; 


* 


* 




;• 


99 






( » 






















# T 


*R 


A 


*N 


*s 


*L 


* 


AT 




*s 


*Y 


ST 


E 


*M 




D 


E 


• • 

V 


E 


• 

L 


• 

0 


# P 


ED 




BY 


TH 








































































t 




































0 








































































N 




























































• • 


• 


• 




• 


• 


• 


M 


• 




• t 






• 


• • 


• • 


• 


• 


• t 




t • 


9 








# 




• • 












• 














































• 














• 














• 


# 




A 


Th 


"m 


A 


*T 


i 


C 


’s 




AN 




* 


A 


*P 


V 


*L 


I 


c 


• 


AT 


’s 




* 


D 


E 




?A 


* 


ME 




it 






IS 
























l> 




















i 














R 




N 






















































0 














i 




T 










fl * 












































N 


































• 


• 


• • 




• 


• 






• 


• • 


• • 


• 


• • 


• 


• • 






• • 


c 












« 






# 




• • 


# 




• 
















• * 


























• 


• • 














• • 








• • 










• 


# 


I 


B 






s? 


ffi 




• 


A 


*M 


ER 


i 


c 


A 


*N 




* 


f p 


# R 


!n 


*T 


IN 




* 


H 


OU 


*s 


E 




FO 


TH 




















T 


E 
































G 














R 


E 




















H 






























































• 


• 








• 






• • 


• 


• 


• • 




• 


• 


• 


• • 


• 


• 


• 


• 














# 


## 




# 




• • • 






























• • 




• • 










• 




• « 
































B 


*L 


: 






TH 

E 




CO 

H 


’p 


*0 


*T 


ER 




*T 


*R 


A 


*N 


’s 


• 

L 


A 


*T 


E 


's 




*P 


*R 


?N 


*T 


ED 




*T 


E 


*X *T 












. 


• 


. 




. 






• 






• 


# 




, 








# 


# 


# 


# 






• • 










# 










• 


* * 




• 


• • 


• 




• 


• 




• 






• 


• • 


•• 


• • 


♦ 


• 






• 


• 


• 




• 


• 


• 




• 




• 










IN 


TO 




e 


*R 


'l 




WH 






LI 




*x 


’o 


*R 


*T 


H 


AN 


, 




*U 


*s 


E 


*s 




•s 


# P 


E 


c 


I 


A 


*L 






• 


















I 






K 














D 




















































c 






E 


































































H 






















































• 




M 


# 


• • 


« 


, 




M 






• • 


, 


• * 


• 


• • 




• 


# 


• • 






• 






















































• 




• 








• 






• • 












• 


• 


• • 
















c 


*0 


0 


E 


•s 




f6 




ci 




*0 


*N 


*L 


• • 

Y 




• • 

U 


*s 


ED 




*L 


*R 




• • 

CO 


B 


IN 


• 


AT 

I 


*S 


I 




























R 




M 




























M 














































































0 








































































N 






















• 






• • 


• 


• 


• • 




• 


• 




• • 


• 


« 


• • 


• 


• • 


• 


9 




• • 


« 


• • 


• 




• 




• 






• 


# 










• 






• 




• • 


•• 






• 




• 


•• 


• 


• 






• 


• 






• 


• 


• 




• 




• 


• 






• 


• 


• 






th 




CO 


*p 


'u 


*T 


ER 




A' 


T 




*P 


*R 


*0 


D 


*u 


C 


E 


's 




C 


AR 


D 


*S 




WH 




AR 


E 




*U 


*S 


ED 


























































C 
















• 
























































H 




















• • 




t # 


*, 


• • 










• • 


• * 


* 


* 






• 




* # 


# * 


I 






# # 


* 




•• 


J 


* 




* 


# * 




H 


# * 


• 




D 


I 


*R 


E 


c 


*T 


*L 


*Y 




TO 




A 


*K 


E 




• • 

TH 




E 


*M 


B 


*0 


*s 


*s 


ED 




*p 


*L 


A 


*T 


E 


*S 




to 


• • 

TH 




































t 


































R 


E 


• 






* 




\ 






* 




* 




M 


* 




• • 




* 


• • 


* 


.* 


, * 


t • 
































* 


B 


*R 


*L 






*u 


B 


*L 


I 


SH 


?fi 




’P 


*R 


*0 


c 


E 


's 


*S 


J 




























• 


























G 

























































































































This sample page was produced by a computer translating English text into braille symbols. The braille equivalent appears di- 
rectly above the English text; this makes it possible to edit the copy without having a special knowledge of braille. 

12 COMPUTERS and AUTOMATION for October, 1959 







This braille embossing machine is linked to a punched card reading device, and makes it possible to produce automatically the metal 
plates used to print braille books. The punched cards containing the braille language are prepared directly from the punched English 

text by an IBM 704 computer at high speeds. 



Texts to be translated by the computer are first trans- 
ferred to punched cards by means of a key punch with 
a conventional typewriter keyboard. The cards are then 
fed directly into the computer, which has had previously 
stored in its memory a program expressing the rules 
for conversion into braille. The machine executes as 
many as 600 instructions per word in less than the 
fortieth part of a second. The computer determines 
contractions and abbreviations by matching against an 
alphabetical table of braille equivalents. 

The translated text emerges from the computer in 
coded symbols on punched cards, and these, in turn, 
are fed to a printer which reproduces the braille symbols 
above the English text for editing purposes. After 
editing, the corrected deck of punched cards is fed to 
an embossing machine, which creates metal plates for 
use in a rotary printing press. 

SIXTY SECONDS FROM ENGINE TESTING 
TO FINAL EVALUATION 
Roger E. Holmes 
Consolidated Electrodynamics Corp. 

Pasadena, Calif. 

The only product of a cell for testing a jet or missile 
engine is data (raw information). Systems of equip- 
ment have been built to acquire a maximum of data from 

COMPUTERS and AUTOMATION for October, 1959 



a test, to convert it into numerical values, and to record 
these values on magnetic tape for eventual input to a 
digital computer. 

At this point, in many cases, engineers have been 
forced to stop; stop short of the goal in rapidly acquir- 
ing this data, which is to improve the performance of 
the engine being tested: they stop because of waiting 
for the computer’s results. 

In order to accelerate the flow of data, a new system 
called a "High-Speed Automatic Data Recording and 
Monitoring System" has been designed and built: it 

processes as well as acquires data from engine test 
cells. A small general-purpose computer is an integral 
part of the system. The system was built by Consoli- 
dated Systems Corp., a subsidiary of Consolidated Elec- 
trodynamics. 

This new system is in use at Pratt and Whitney’s 
Florida Research and Development Center, and the 
entire operation from acquisition of data from an en- 
gine test to final computed results takes only 60 seconds. 
As a result, research and development engineers can 
make immediate decisions from the final results even 
while the engine is still running. Accuracy and flexi- 
bility are other advantages of including the computer 
in the system. 






MORE ADEQUATE SELECTION OF CANDIDATES 
FOR JOBS 

An electronic data processing system has been installed 
in Randolph Air Force Base, Texas, for the Air Training 
Command, with its main assignment the more adequate 
selection of personnel. The computer has been pro- 
grammed by the ATC Statistical Services group for this 
purpose. Screening of hundreds of records can now be 
accomplished in minutes. 

For example, three people may need to be selected for 
assignment to a certain job. ATC personnel "asks” 
the computer for the 10 people in the command most 
qualified for such a position. The information necessary 
to answer this question, consisting of records for each 
officer, airman, and civilian in the command — are 
maintained on magnetic tapes which store assignment 
preferences, additional and potential skills, technical, 
flying, or professional training completed, and many 
other types of information. Each individual record is 
updated daily. 

From this store of knowledge, the computer, a Bur- 
roughs Datatron 220, seeks 10 people eligible for the 
job ATC Personnel desires to fill. In a short time, 10 
names are produced, together with their pertinent back- 
ground. A selection board of human beings may then be 
designated to select three people out of the 10 for the 
assignment. 

Much the same program may be used to fill promo- 
tions, quotas for overseas duty, assignments to schooling, 
etc. In addition to these personnel operations, the com- 
puter will soon be given other jobs, such as mainte- 
nance of records on aircraft, materiel, and vehicles. 

The 220 at ATC is one of the largest of the Datatrons 
yet installed. It includes a 100,000-digit magnetic core 
memory, and eight magnetic tape auxiliary storage 
units, with a total storage capacity of approximately 
112 million digits. 

ALL DAILY WORK FOR BANK CHECK 
ACCOUNTING 

All daily bookkeeping tasks for 220,000 checking ac- 
counts at 60 branches of the Bank of America in the Los 
Angeles, Calif., area have gone on to electronic data pro- 
cessing machinery, as of September, 1959. The system be- 
ing used is the production model of ERMA, now called the 
GE 150 and GE 210 systems (ERMA is an abbreviation 
for Electronic Recording Method of Accounting). The 
system uses magnetic character reading. It is manufactured 
by the General Electric Computer Department, Phoenix, 
Arizona. 

The system reads checks, sorts them, and posts 550 ac- 
counts per minute. By comparison, an efficient bookkeeper 
with a year’s experience, can sort and post to about 4 ac- 
counts per minute. 

The system processes and records all of today’s checking 
transactions before tomorrow’s banking begins. The Bank 
of America expects that at least 13 ERMA centers will be 
operating in California by the end of 1961, serving its 463 
branches, and more than 2 million accounts. 

ERMA comprises essentially a document-handler, an 
electronic computer which controls the entire system, a 
control console, tape units for storing information, a printer, 
and other input and output peripheral equipment. The 



printer, through directions from the computer, prepares 
customer and branch-bank statements at the rate of 900 
lines per minute, or about 50 complete statements per 
minute. 

The system is completely transistorized. The magnetic 
character reader can handle and read folded, mutilated, 
spindled, and overstamped checks and deposit slips. 

The checks used by the Bank of America are similar to 
conventional checks, with one exception — stylized numbers 
are imprinted in magnetic ink across the lower edge. 
Numbers are in Arabic numerals, running from 0 through 
9, and can be read both visually and by ERMA. They are 
imprinted in a special type font, developed by General 
Electric and standardized by the American Banking Associa- 
tion. Code numbers represent the issuing bank’s number, 
routing symbol, and identifying data such as branch num- 
ber and account number. 

The system operates as follows: 

1. The customer writes a check which has been pre- 
printed in magnetic ink with ABA number and account 
number; 

2. The check is presented to the bank, and the dollar 
amount is registered by a bank clerk in magnetic ink during 
the proofing operation; 

3. The check is fed through the document handler, read 
by the magnetic character reader, and the information 
transmitted to the computer. 

4. The computer performs some checking of the infor- 
mation, and sorts at a speed of 30,000 decimal digits per 
second, recording the information from the checks in ac- 
count-number sequence; 

5. The sorted information is compared with information 
in the customer’s record, and the new account balance and 
the check amount are recorded in the customer’s file; 

6. While the customer account is being posted the 
check is sorted off-line on the document handler and goes 
to a customer file; 

7. The high-speed printer produces a statement for the 
customer, showing all transactions and the current balance, 
either at the end of month or upon request; 

8. As ERMA-sorted checks are filed, the signatures are 
compared with checks previously filed. 

9. The cancelled checks and statements are sent to the 
customer, usually once a month. 

SANTE FE RAILWAY RESERVATION SYSTEM 

The process of obtaining a railroad reservation has 
long been an irritating and frustrating part of railroad 
travel. This problem has been a source of concern and 
the subject of study by many of the country’s leading 
railroads, but relief is now at hand through the help of 
modern electronics. 

In July the Atchison, Topeka and Santa Fe Railway 
put into service an automated electronic system for han- 
dling its passenger reservations. This combination data 
processing and communications system enables Santa 
Fe reservations personnel from Los Angeles to Chicago 
to make reservations for customers in a matter of sec- 
onds instead of hours or even days as has been the case 
in the past. The system is called the Magnetronic Res- 
ervisor and was designed, built, installed, and is main- 
tained by The Teleregister Corporation of Stamford, 
Connecticut. 



14 



COMPUTERS and AUTOMATION for October, 1959 





This "push-button” train reservation agent’s set automatically clears and reserves Pullman space and chair car seats in a matter of 
seconds. It is in operation in five major cities served by the Santa Fe Railway: Chicago; Kansas City, Mo.; Fort Worth, Tex.; 
Los Angeles; and San Francisco. The agent’s set can handle eleven separate types of transactions. 



This equipment will process for Santa Fe Railway’s 
sales agents all standard railroad accommodation trans- 
actions: (1) reserving the accommodation; (2) selling 
the accommodation and removing it from the inventory ; 
and (3) cancelling previously made reservations when 
necessary. Agents perform these transactions through a 
piece of equipment not much larger than a typewriter, 
which produces a duplicate printed record of each trans- 
action including a complete identification of the train 
space involved. This information includes the date, the 
train number, the car number, the particular seat, draw- 
ing room or bedroom number, and a customer’s identi- 
fication number. 

Each of the Santa Fe agent sets is connected through 
specialized communications equipment and over a net- 
work of leased lines to the center of the Magnetronic 
Reservisor which is located at LaSalle Street Station, 
Chicago. This is the data processing center for all of 
Santa Fe’s reservations information. It consists pri- 
marily of two magnetic memory drums and some forty- 
odd racks of computing and communications equipment. 
This is capable of keeping track of a total of 148,000 
accommodations for any period up to 6 months prior 
to train departure date. This equipment is also capable 
of handling a reservations transaction originated in one 
of the agent sets in approximately a half second. 

In addition to the standard accommodations transac- 
tions, the system provides the Santa Fe with the fol- 
lowing services : 

COMPUTERS and AUTOMATION for October, 1959 



a. Cushion Control — provides advance warning when 
available space is critically low. 

b. Priority Control — provides for systematic selling 
of available space, in response to random requests, at 
any time and in any sequence desired by management. 

c. Wait-list Control — keeps a record of requests for 
which there is no space available so that the first can- 
cellation can be given to the first person on the waiting 
list. 

d. Expiration and Cancellation Control — When the 
agent reserves space, he informs the customer that the 
ticket must be picked up and paid for by a certain date 
and time. In the event that the customer fails to pick 
up the ticket before this expiration time, the Reservisor 
automatically cancels the reservation, making it available 
for another passenger. 

e. Records Display — provides for printed records 
of any of the sales information or data stored in the 
machine to be quickly available to railroad management. 
This information is automatically arranged by the ma- 
chine in the most convenient and useful manner, such 
as passenger lists for conductors (called consists), a 
count or list of transactions for each train car, a count 
of all the transactions by each agent, and many other im- 
portant reports and statistics. 

The Reservisor is in service seven days a week, 365 
days a year, and is only taken off line for testing and 
maintenance routine during certain hours in the early 
morning when there is virtually no traffic. The data 



15 




processing equipment is designed for rapid and easy 
maintenance by the use of plug-in elements which, when 
they require replacement, may be removed and changed 
in a few seconds. Dual facilities are provided for critical 
functions so that one piece of equipment constantly 
double-checks the performance of its opposite number; 
also, if one rack has to be taken off line for mainte- 
nance, its dual counterpart carries on the functions 
by itself. 

CASUALTY INSURANCE ACTUARIAL 
CALCULATIONS 

An electronic computer in use in the National Bureau 
of Casualty Underwriters, New York City, is devoted al- 
most entirely to actuarial calculations — the development 
of rate revisions, the development of actuarial tables, and 
research problems. The Bureau serves as a statistical and 
rate-determining agency for casualty insurance companies 
located throughout the country. The computer is the 300th 
LGP-30 put into operation since the equipment was first 
marketed in 1956 by Royal McBee Corp. 

More than forty million punched cards are received 
annually in the Bureau from member and subscriber 
companies, reporting premium and loss data. The ex- 
perience is checked, sorted, and summarized by sta- 
tistical categories such as year of coverage, class, and 
territory, so that a meaningful analysis can be made. The 
electronic computer analyzes this summarized experience. 

For example, in revising liability rates on private passen- 
ger automobiles, the computer determines trend factors, 
rate of level changes by territory, final rates by classification, 
and a comparison of present and proposed rates. 

In the case of the trend factor calculations, the input 
data consists of the average claim costs for five years for 
each state, which represents five points on a trend line. A 
line of best fit is computed by the least squares method and 
projected to 18, 21, and 24 months in the future. Trend 
factors, equal to the projected figures divided by the latest 
year’s least square value, are computed for each state and 
the country as a whole. The trend factors represent the 
estimated increase or decrease in average claim costs in 
the future months when new proposed rates will be in 
effect. 

RAY TRACING IN A CASSEGRAINIAN 
OPTICAL SYSTEM 

T. Sabine 

Avco Corp., Crosley Div. 

Cincinnati 25, Ohio 

A cassegrainian optical system consists of two re- 
flectors arranged in such a way that the secondary or 
hyperbolic reflector intercepts the rays coming from 
the primary or parabolic reflector before they reach the 
focus. These rays are then sent through a hole in the 
primary reflector to the focal plane of the combination. 

The problem is to determine the best geometry for 
such a system: the optimum sizes and focal lengths of 
the reflectors. 

In a recent case of ours, the project manager for the 
development of this system placed limiting restrictions 
on the following items: 

1) Aperture of the primary reflector; 

2) Maximum overall size of the container; 



3) Off-axis image size. 

The method to be used was to trace extreme light 
rays through the double reflection to the focal plane, 
using systematically varied values for focal length, spac- 
ing and the diameter of the secondary reflector. The 
combination of paraboloid primary and hyperboloid 
secondary is exactly right for sharp focus on the axis, 
as a person who has studied conic sections knows. But 
the problem was to compute the total off-axis aberration 
as a function of the system geometry. This computation 
has been made in the literature of reflecting telescope 
design, with the results of suggesting some fourth-degree 
curves for the mirrors (Ritchey or Schwartzschild) 
which improve off-axis performance enormously at small 
cost in axial image blurring. But the number-juggling 
ability of a computer is such that the easiest approach 
to the design figures wanted here was to trace the rays 
to intersection with the secondary reflector, compute 
incidence (and therefore reflection) angles and then 
compute points of intersection with the focal plane, just 
as if the problem had never been considered before. 

This was done by computing components of the di- 
rection vector v of a light ray striking an optical surface 
whose normal direction vector is n from an original di- 
rection vector u. Thus the machine did hundreds of 
calculations of the type v=:u — 2u • (nn), where nn 
is a dyadic which is a special type of tensor. 

Location of reflection points was simply a matter of 
solid analytic geometry, which the IBM 650 handles 
with ease. If a ray starts at some point Pj (x-,, y x , z x ) 
and intersects the secondary reflector at P 2 (x 2 , y 2 , z 2 ) 
then the machine resolves compatibility between: 
x 2 /a 2 — (y 2 + z 2 )/b 2 = 1 

and d = (x 2 — x^/x' = (y 2 - yi)/y' = (z 2 - z x )/z' 
where d is the distance from Pj to P 2 and x', y', z' are 
direction cosines of the ray direction. Similar compu- 
tation locates the point of intersection with the focal 
plane. A sufficiently large number of points P x had to 
be used for each mirror geometry to be a good sample 
of primary coverage since it is not necessarily true that 
the extreme incident rays end up defining the boundaries 
of the focal spot. 

Actually, many test cases were computed and by using 
the results obtained from one test, a new and better 
guess was made for the next test, until a result was ob- 
tained well within the limiting specifications. 

The mathematical and physical principles involved in 
this problem were not difficult; however, the volume 
of computations necessary to define this system was very 
large and this method could not have been used if it had 
not been for the powers of the digital computer. 



AUTOMATIC MACHINING OF PARTS WITH 
TWO-AND THREE-DIMENSIONAL CONTOURS 

Robert F. Heslen 

Bendix Aviation Corp., Industrial Controls Section 
Detroit 37, Mich. 

A numerical-control tape-preparation system has been 
operating which automatically prepares control tapes for 
two- and three-dimensional contour machining of parts 
composed of straight lines, arcs, circles, and free form or 
point defined curves. 



16 



COMPUTERS and AUTOMATION for October, 1959 




This system is a digital computer system designed es- 
pecially for numerically controlled machining operations. 
It takes in dimensional and machining information from 
engineering drawings, and puts out a coded punched con- 
trol tape for direct control of machine tools. This system 
has been used to prepare machine control tapes for machin- 
ing intricate three dimensional parts containing pockets, 
flanges, tapers, contours, and grooves. Such parts include 
turbine blades, intricate cams, molds, forging and stamp- 
ing dies, templates, integral panels, forged fittings, etc. 

The equipment required in this system for producing a 
control tape includes: a prepared form, caled a process 
sheet; a flexo-writer with a tape punch; a Bendix G-15D 
general purpose digital computer; and a set of computer 
programs. 

Computer Program for Automatic Control 

Starting with a drawing of the part which has all dimen- 
sions given in the form of conventional rectangular co- 
ordinates referenced to a set of axes, or in some cases, a 
mathematical description, a process sheet is prepared. 

The process sheet lists the coordinate dimensions of the 
part in the desired cutting sequence, along with such infor- 
mation as machine feed rates, cutter diameter, and toler- 
ance. Coordinates at the end of each straight line section 
must be entered. For circles and acres of circles, only radii 
and end points of arcs need be specified. The computer 
will approximate the arcs by a series of straight lines 
(chords) to the accuracy specified in the tolerance entry. 

The process sheet is then reproduced on a special type- 
writer ( Flexowriter ) which simultaneously prepares a 
punched process tape and condenses information from the 
process sheet into a form suitable for use as input to a com- 
puter. Verification of process tape accuracy is accomplished 
by inserting it in a tape reader unit attached to the type- 
writer. A second transcription of the process sheet is then 
made to produce a second process tape. If the two codes 
compare, the tape punch will operate and punch a verified 
second process tape. If the codes do not compare, the 
punch will not operate and the operator can then locate 
the error. Thus, the second tape is verified tape and will 
not contain any transcription errors. The verified process 
tape is then inserted into a reader that operates in conjunc- 
tion with the computer. 

The process tape is read by the computer which auto- 
matically produces the control tape in the form required 
for machine control. 

The computer will: (1) calculate cutter center loci, (2) 
interpolate curve sections to provide intermediate points 
defining a continuous cutter path, (3) resolve specified 
feed rate into required speeds of the individual machine 
slides, (4) translate decimal input information into binary 
coded form, and ( 5 ) generate parity checking information. 
The interpolation function performed by the computer es- 
tablishes the required spacing of intermediate points on 
curved sections in accordance with the tolerance specified 
on the process sheet. 

Since the cutter path between the sequential points on 
the control tape will be along a straight line, the spacing 
of points will vary continuously so that the deviation from 
the desired curve will always be sufficiently small to satisfy 
the tolerance and surface finish requirements. 

A tooling or process engineer can be trained to use this 
system in less than a week. 

COMPUTERS and AUTOMATION for October, 1959 



PREDICTING THE SOUNDNESS OF FOUNDRY 
CASTINGS 

A. A. B. Pritsker 

Battelle Memorial Inst. 

Columbus, Ohio 

Of continuous concern to foundrymen is their ability to 
make castings free of defects such as shrink cavities. In 
the past, sound castings have usually been obtained by ap- 
plication of empirical techniques. However, the use of 
this approach has become more difficult as the casting art 
has been asked to deal with more complex situations; for 
example, the casting of high-purity metals requires en- 
closure in a vacuum tank, and in this case it is difficult to 
use such techniques as "pour-out” in order to determine the 
location of the solidification front at various times — the 
movement of this front determine whether or not shrink 
cavities will form. 

Now a solidifying casting changes its state from liquid 
to solid by losing heat. Therefore analysis of the move- 
ment of the solidification front is possible by determining 
the heat flow, and the resulting isothermal patterns in the 
casting at succeeding time intervals. Previous analysis of 
this problem has applied classical techniques. But these 
were necessarily restricted in scope because of the simplifi- 
cations required in order to obtain a solution in closed 
form. Passive network analyzers have also been used, but 
the complexity of the problem requires the use of an un- 
desirably great number of high-precision components in 
the experiment. 

By representing a casting and the associated mold with 
a series of small cells, it has been found possible to use 
finite-difference techniques to describe the heat flow, and 
the resulting solidification patterns, in a casting. This 
technique represents an extension of previously used man- 
ual methods, which were however severely restrictd in ap- 
plication because of the length of time required to obtain 
a solution. By using an IBM 650 computer it has been found 
feasible to include many factors that were previously ignored 
and at the same time cut down the length of time required 
for a solution. This has made it possible to obtain solu- 
tions that have shown excellent correlation with experi- 
mental work conducted concurrently to check the validity 
of the mathematical model. 

By using the model and the computer it is now possible 
to predict the effects of changes in casting procedures upon 
the soundness of the casting, and this analysis is being used 
to replace costly experiments. 

The present model is regarded as a first step in the use 
of digital computers as an aid to foundrymen. It is ex- 
pected that more sophisticated models will be developed 
for more complex studies of foundry castings. 

COMPUTERS SIMULATING MEN 
F. Massnick 

Minneapolis-Honeywell Aeronautical Division 
Minneapolis, Minn. 

Both men and aircraft have been simulated on digital 
and analog computers by Minneapolis-Honeywell in studies 
to determine human ability to control devices ranging 
from a single push-button to a complete space ship. 

The computer technique may be termed "robot psycholo- 
gy”; the use of computers has speeded up exploratory re- 
search by eliminating the variability of real men. 

17 




The most important aspects of controls which men use 
appear to be those concerned with the dynamics of the 
control and of the control display combination rather than 
the particular location, shape and size. 

Some of the study has been devoted to human abilities 
in the control of aircraft. 

Although a great deal of prior research has been carried 
out in the study of human behaviour, many of the results 
are not in a form adapted to the design of actual systems 
and control devices. So a large part of the current study 
has been to examine, classify and evaluate data, and try to 
bring them into a coherent form leading to a general solu- 
tion of the problem of human control. 

Among the studies which have been carried out was an 
experiment on the behaviour of systems guided by "model” 
pilots. Both the man and the aircraft were simulated on a 
computer and the effect of variation in both human and 
machine parameters measured. 

Although this particular kind of "psychology” deals with 
synthetic men rather than real men, it can produce results 
which apply to the control of aircraft by real men. 

Experiments can be done rapidly and without the varia- 
bility introduced by real subjects. After such "robot psy- 
chology” has been worked out, the limits are known within 
which it is likely to be profitable to study the behaviour 
of real men. 

FLOOD CONTROL BY COMPUTER 

The U. S. Army Corps of Engineers is now assessing 
rainfall electronically. By using a computer for one of the 
nation’s flood-control areas, the Ohio River Basin, hydrolo- 
gists can take in rainfall data, predict rising water levels, 
and calculate the re-routing of water from over filled 
tributaries. 

The Corps of Engineers’ Ohio River Division, based in 
Cincinnati, calculates a typical flood routing problem in 
five minutes with the E101 Burroughs computer, whereas 
prior methods would require 75 minutes. 

Flood routing is the procedure used to advance a flood 
wave downstream from point to point, modifying it by 
the effects of the channel and addition of tributary inflow. 

Computational methods for routing vary according to 
the basin. A process known as the Muskingum Method is 
employed in the Ohio River Basin. Using a simple mathe- 
matical formula, the computer determines when a flood, 
either known or predicted, can be routed to downstream 
points for maximum dissipation. 

Other computations are used by hydrologists to convert 
rainfall over an area into flow in the stream draining that 
area. It is, in effect, a means of determining from the cur- 
rent season’s rainfall, and from historic measurements, how 
high the waterflow will be. 

For example, if through their computations hydraulic en- 
gineers find that the March rainfall may produce rivers 
swollen enough to produce flood conditions, the informa- 
tion is relayed to flood control experts who take preventive 
measures — re-routing, banking, etc. 

AUTOMATICALLY 
SHAPING NONCIRCULAR GEARS 

Cunningham Industries, Inc., Stamford, Connecticut 
has developed a program for the Royal Precision LGP- 
30, marketed by the Royal McBee Corporation, which 
makes possible automatic design and production of non- 
circular gears. 

18 



Input consists of coordinates of points describing a 
mathematical function. Output is a punched paper tape 
which can be used as direct input to a gear shaper to 
control automatically the shaping of a pair of gears. 

No separate equipment for input data tape prepara- 
tion is needed since a modified electric typewriter that 
produces punched paper tape is part of the LGP-30 
computer. 

On more limited computing equipment, the solution 
of the whole problem required four programs whereas 
with the LGP-30 it requires only one program. 

RESERVOIR FILLING 

An IBM 650 computer was used by the Colorado 
River Board of California to produce in 56 seconds a 
complicated water and power study over 36 years of 
observations that would have taken three engineers more 
than six months to compute with "paper and pencil’’ 
methods. The need for the computer’s study of the 
river flows and other important data was created by the 
building of the gigantic Glen Canyon reservoir on the 
Colorado River. The 650 study was used by the Board 
in determining the best way to fill the 180-mile-long 
reservoir for the dam so that water and power supplies 
for Southern California, Arizona, and Nevada would 
not be adversely affected. 

SEA RESCUE OPERATIONS 

A dramatic application of a RAMAC 305 has been 
made by the U.S. Coast Guard. This machine is installed 
in the Eastern Area Command headquarters in New 
York City, and is being used to expedite search and 
rescue operations in the North Atlantic Maritime Re- 
gion, which also includes the Caribbean Sea and the 
Gulf of Mexico. Each day the computer charts the 
course for all merchant vessels in this area, regardless 
of nationality. In case of emergency, the latitude and 
longitude of the ship in distress is entered in the com- 
puter. The IBM RAMAC 305 will then automatically 
determine which ships are in the immediately surround- 
ing areas, so that the best situated ships can be con- 
tacted and requested to participate in rescue operations. 

WEATHER FORECASTING 

An IBM 705 in Asheville, N.C., makes available long- 
range climatological studies. Vast quantities of weather 
information are condensed on coded tapes and stored 
in the machine’s memory. When a particular weather 
problem for some area comes up, a run on the 705 
will quickly deliver a detailed analysis of the problem. 
The computer has been used to plan many long-distance 
military flights. 

At Suitland, Maryland, the Weather Bureau in co- 
operation with the Air Force and the Navy has per- 
fected a program for computing the weather on an IBM 
704, based upon radio reports from around the world. 
In devising this program and checking its results, many 
long-held theories about the weather were found to 
have no real effect on the computer’s predictions, and 
therefore have been dropped or changed. The present 
weather map, which is printed daily by the 704, is con- 
sidered to be 15% more accurate than any map weather 
map made by human beings, and is now the only 
weather map used by the Air Force and - Navy for 
weather forecasting. 

COMPUTERS and AUTOMATION for October, 195? 





THE HUMAN FACTOR 



in today’s technology 



Scientists have long been preoccupied with the techno- 
logical problems of Man and the Machine. The increas- 
ingly complex nature of advanced systems has created 
an urgent need to enhance man’s contribution to effective 
systems performance. The complicated nature of this 
relationship requires the skills of psychologists, social 
scientists, mathematicians, and engineers. 

At Ramo -Wooldridge, human engineering, personnel 
selection, individual and system training, display design, 
and communications are successfully integrated into 
systems design and development by the technique of 
large-scale simulation. 



Simulated inputs enable scientists to observe a system as 
it operates in a controlled environment and make possible 
the collection of data on performance, training, human 
engineering, maintenance, and logistics and support. 
Scientists and engineers use this data to assure the design, 
production, and delivery of a unified system capable of 
high performance and reliability. 

Expanding programs at Ramo -Wooldridge in the broad 
areas of electronic systems technology, computers, and 
data processing have created outstanding opportunities 
for scientists and engineers. For further information con- 
cerning these opportunities write to Mr. D. L. Pyke. 






RAMO -WOOLDRIDGE 

P. O. BOX 90534, AIRPORT STATION • LOS ANGELES 45, CALIFORNIA 

© I a division of Thompson Ramo Wooldridge Inc. 



COMPUTERS find AUTOMATION for October, 195 9 



19 




Machine Translation Methods, And 
Their Application To Translation From 

English To Russian 

I. K. Belskaya 

Academy of the Sciences of the U.S.S.R. 

Moscow, U.S.S.R. 

(Based on a portion of a paper given at the International Conference on Information Processing, 

Paris, France, June, 1959) 



R ESEARCH in methods of machine translation was 
started in the Soviet Union late in 1954 on the ini- 
tiative of Academician A. S. Nesmejanov, President of 
the USSR Academy of Sciences. The first experiments 
in machine translation (MT) from English into Rus- 
sian were carried out in December, 1955. 

Since then considerable progress has been made to- 
wards adequate formulation of the method to be used. 
We can now say that a second stage of the research has 
recently been completed showing that the suggested 
methods are of general applicability. For demonstrating 
this, these methods have been extended to cover MT 
from languages differing greatly from English in struc- 
ture, such as Japanese, Chinese and German. 

In regard to the MT methods for translating from 
English to Russian, the research has reached a third 
stage in which: (1) complete grammatical analysis on a 
bilingual level takes place; (2) rearrangement of the 
most important types of English idiomatic constructions 
is accomplished; (3) grammatical modification of the 
Russian translation is performed by an independent set 
of routines, which have been termed Russian Synthesis. 

In addition, in the field of applied mathematics, the 
volume of words now entered into the MT dictionary 
has considerably increased. More than 2000 words are 
stored in the English section of the multilingual MT dic- 
tionary; a still greater number of Russian equivalents is 
stored in its Russian section. The dictionary thus covers 
many different areas of applied mathematics. 

To complete this stage of research a large-scale experi- 
ment on the Anglo-Russian MT methods has been car- 
ried out. One hundred samples (which amounted to 
3000 sentences) of "unknown text” were selected at ran- 
dom from different English authors, and translated man- 
ually into Russian in strict accordance with the instruc- 
tions provided by the MT dictionary and the translating 
routines. The ten persons chosen to carry out the ex- 
periment had no knowledge of English nor had they 
any previous experience with the task required. 

This experiment showed that the scheme appears to 
be very effective at dealing with all sorts of texts so long 
as they are restricted, lexically, to applied mathematics. 
Grammatically no limitation as to type of the written 
text has been found necessary. One or two words per 
printed page is the average for "unknown” words with 

20 



the present-size dictionary. This makes the translation 
quite adequate for understanding. 

For this reason, as well as for reasons of preserving 
the proposed series of MT dictionaries strictly specialized 
as to field, we are not inclined to increase the volume of 
words in the present dictionary, but intend rather to 
proceed with compiling medium-sized (say, 2500-3000 
words each) dictionaries for various fields. This indeed 
will be our occupation at the next stage of research. 

Translating routines for Anglo-Russian MT being the 
final achievement of the recent research, it seems reason- 
able in the present report, to lay particular stress on a 
description of translating routines for vocabulary and 
grammatical analysis of the English sentence. 

Applicability of MT Methods 

The two most general MT questions are: is it possi- 
ble to translate by machine? is it practical? The former 
question has already been answered, both theoretically 
and practically. The latter question still remains open 
for discussion. The objective of our present stage of re- 
search is to prove the applicability, the practicality of 
MT methods to any sphere of language. 

To date, only within the limited sphere of scientific 
writing has the applicability of MT methods won gen- 
eral recognition. As to other uses of MT, most machine 
translators are inclined to feel very doubtful. 

However, the majority of restrictions imposed on MT 
application, when analyzed, turn out to be due to a very 
strong inclination on the part of investigators to des- 
cribe the translated language (source language) in terms 
of correspondence with some other system, say, another 
language, or a group of languages, or a science other 
than linguistics, especially particular logics or particular 
fields in mathematics. The possibilities of MT are dis- 
cussed then as dependent on common elements in the 
compared systems. These elements may be more or 
less numerous, yet absence of complete correspondence 
between the systems, which is usually the case, inevitably 
brings about limitations of the scope of MT. Thus, ap- 
plication of machines to translating literary works of art 
has more than once been declared to be absolutely ruled 
out. 

[Please turn to page 22} 

COMPUTERS and AUTOMATION for October, 1959 





Space Technology Laboratories' new corporate symbol represents a bright history in a stimulating age. stl has 
provided the over-all systems engineering and technical direction for the Air Force Ballistic Missile Program since it 
was assigned the highest national priority in 1954 . Five years of accelerated effort produced epic advances in science 
and technology, and propelled the art of missilery through three distinct generations of progress. STL contributed 
technical leadership to the science / government / industry team which has built this solid , expandable foundation 
for future advances in space , and is daily adding new strength to our national security. In addition to its major 
management functions, STL also conducts advanced space probe experiments for the Air Force at the direction of 
such agencies as NASA and ARPA. ^ To those scientists and engineers § with capabilities in propulsion, electronics, 
thermodynamics , aerodynamics , structures, astrophysics, computer / technology, and other related fields and 
disciplines, stl now offers unique professional opportunities. Inquiries A regarding staff positions at STL are invited. 




a new symbol 
for a new era of 
technology 



Space Technology# Laboratories, Inc. P. 0. BOX 95004, LOS ANGELES 45, CALIFORNIA 

COMPUTERS and AUTOMATION for October, 1959 2: 





[Continued from page 20] 

In our opinion, it seems very reasonable to expect that 
these limitations can easily be eliminated, if the problem 
is formulated in a different way: namely, is it possible, 
within any existing language, to give formal description 
to any of its multiple spheres of use, individual as they 
may seem? 

This is the same question as asking if the applicability 
of MT depends on whether it is possible to identify the 
implicit set of rules governing this or that particular 
sphere of language — perhaps as narrow a sphere as, say, 
Wordsworth’s poetry, and, asking whether these rules 
can be formulated into a formal set. 

It is our opinion that every piece of writing (insofar 
as written language is discussed) can be analy 2 ed along 
these lines within the sphere where it belongs, and a set 
of rules for such analysis can be laid out. It is essential 
that these rules should be formal everywhere. Yet this 
is no obstacle either, since language is but a formal sys- 
tem of specific character developed by human beings to 
express and communicate their mental activities. If this 
is true, problems posed by stylistic peculiarities of liter- 
ary works of art can be satisfactorily resolved, if treated 
along the lines suggested above, i.e., within the sphere 
where they belong. 

In this light, the supposed "principal informaliza- 
bility” of poetry should be rejected. On the contrary, 
poetry, as indeed any piece of literary art where formal 
elements are of no minor importance, would be particu- 
larly susceptible to machine translation in this sense. 

Our opinion has been partly justified on empirical evi- 
dence, that is, by experimental translation of passages 
from Charles Dickens, J. Galsworthy, J. Aldridge, and 
Edgar A. Poe. It is our firm belief that further investi- 
gations will completely eliminate the restrictions now 
generally thought to cover MT application. 

Finally, an adequate description of a language, or of 
any particular sphere of it, should aim at establishing 
within the language or sphere, a set of correlations of 
linguistic means with linguistic effects, i.e., of words 
and other linguistic devices with their meanings, their 
content, their denotation. 

Taken in its most general sense, the translating prob- 
lem is then the problem of equating these correlations 
in one language with those in another. 

It seems to us that MT prerequisites do not rest upon 
the existence of common basic elements in languages, as 
is often pointed out, but rather include the following 
two factors: 

(1) Language in itself is only a system of formal 
means by which communication of meaning is effected; 

(2) All existing language systems are so developed 
as to express in their particular ways various shades of 
meaning as well as various emotional effects. 

To refer again to our correlation, this is the same as 
asserting that the number of "effects” in any two langu- 
ages is equal. This assertion makes the corresponding 
systems of "means” fully comparable, through their 
"effects.” 

Since language systems are formal, any application of 
them can be provided with a description programmable 
on a machine. 

A Short Outline of the Translating Routines 

The general procedure covered by the translating 

22 



routines we have worked out can be classified into three 
independent steps, these being: 

1. Vocabulary Analysis of the source language, for 
which purpose the MT dictionary and a set of 
Dictionary routines are used ; 

2. Grammatical Analysis of the source language, for 
which purpose Analysis routines are devised ; 

3. Grammatical Synthesis of the target language, for 
which purpose the same set of Synthesis routines 
is applied to the texts translated from different 
source languages. 

To make the outline concrete, the translating routines 
will be further described in their Anglo-Russian realiza- 
tion. 

Dictionary Analysis 

Dictionary Analysis of the English sentence starts with 
searching in the MT dictionary for every word of the 
text. The first dictionary routine to be used here is that 
of transforming words of the text into the standard 
forms listed in the MT dictionary. 

Thus "wanted” will be transformed into "want,” 
"stopped” into "stop,” "coming” into "come,” "lying” 
into "lie,” "copies” into "copy,” "bigger” into "big,” 
etc. 

When the dictionary search is completed, another rou- 
tine is applied which deals with the words that for vari- 
ous reasons have not been found in the dictionary. These 
are termed, "unknown words,” because their lexical 
equivalents remain unknown throughout the translating 
procedure. Yet, for the "grammatical analysis” routine 
to be applied later, grammatical qualification of the "un- 
known words,” must be obtained. 

It is impossible to foresee every word in every text of 
a language or even in one of its particular spheres, since 
some of the words may be occurring for the first time in 
the language, not to mention quite a number of more 
trivial reasons. 

However, the "unknown words” do not affect the 
translation greatly, if they have been classified grammati- 
cally. To meet the latter problem, a very important rou- 
tine, that of classifying "unknown words” into "parts of 
speech” has been devised. In this routine extensive use is 
made of the morphology and syntax of these words. 

Another category of sentence constituents which un- 
dergo preliminary grammatical analysis in accordance 
with a dictionary routine, are the so-called "formulas,” 
by which we mean various symbols used in different 
sciences. The syntactical function of every "formula” in 
the sentence is defined in accordance with a special rou- 
tine. 

So much for the words and symbols not found in the 
MT dictionary. 

In addition to lexical equivalents, words found in the 
dictionary are provided with information (termed "in- 
variant characteristics”) which is partly grammatical and 
partly semantic. The "invariant characteristics” obtained 
from the dictionary distinguish between final and pre- 
liminary information. Information is considered final 
for the dictionary routine when the lexical equivalent 
of the word is included. Information is preliminary when 
it is restricted to the indication "homonymous” or "poly- 
semantic.” 

[Please turn to page 27] 

COMPUTERS and AUTOMATION for October, 1959 






and A U T O M AT I ON 



DATA PROCESSING • CYBERNETICS • ROBOTS 




■■■■ 



in the multi-billion 
dollar computer and 
data processing 

field! A 



nuuuuuu^ 



Perhaps the Key Reason Why 

COMPUTERS AND AUTOMATION 

Is Read By The PEOPLE WHO COUNT 
Is That It Was Founded And Is 
Edited By Edmund C. Berkeley — 
Pioneer Computer Mathematician 



COMPUTERS AND AUTOMATION 

Is The Only Monthly Trade Publica- 
tion Serving Both Builders And 
Buyers In This Mushrooming Multi - 
Billion Dollar Industry 



EDITOR'S 

BACKGROUND 

| 

• Author of "Giant Brains or Machines 

that Think”, John Wiley and Sons, 1949 j 

• Co-author of "Computers — Their 

Operation and Applications”, Reinhold j 

Publishing Co., 1956 j 

• Author of "Symbolic Logic and Intelligent 
Machines”, Reinhold Publishing Co., 1959 

I 

• Maker and Designer of small computing 
machines including the Brainiac® electric 
brain construction kits, Simon (miniature 
complete automatic digital computer), 

Relay Moe (tit-tat-toe playing machine 
pictured in Life Magazine, March 19, 

1956), etc. 

• Fellow of the Society of Actuaries; 

Harvard 1930 A.B., summa cum laude in 
mathematics; author of many articles and 
papers in New York Times Magazine, 

Scientific American, Record of the 
American Institute of Actuaries, Journal 

of Symbolic Logic, etc. I 

; 

• Entered computer field in 1938. 

i 

• Founded "Computers and Automation” 

in 1951. ] 

The Computer & Data Processing 
Field Is Huge! j 

When Edmund C. Berkeley founded 

COMPUTERS and AUTOMATION in 

1951, the industry was just starting to flex j 

its muscles! Now it is estimated that the 
computer and data processing market will 
be "between five billion and nine billion 
dollars over the next five years". j 



1. Computer BUILDER Subscribers include: 



Addressograph Multigraph, Aladdin Electronics, Alwac, 
American Machine and Foundry, Ampex, Arnold 
Engineering, Automatic Electric, Beckman Instruments, 
Bendix Aviation, Bristol Company, Bryant Gage and 
Spindle, Burroughs, Clary Multiplier, Computer Control, 
Datamatic, Daystrom Systems, Electronic Associates, 
Epsco, Ferranti Electric, Ford Instrument, General 
Controls, General Electric, General Kinetics, Goodyear, 
Haller Raymond and Brown, International Business 
Machines, Jet Propulsion Labs, Lear, Leeds & Northrup,. 
Link Aviation, Litton Industries, Melpar, Minneapolis- 
Honeywell, National Cash Register, Northrop Aircraft, 
Packard Bell, Panellit, George A. Philbrick Researches, 
Philco, Potter Instrument, Ramo- Wooldridge, Raytheon 
Mfg., Reeves Instrument, Royal McBee, Sperry-Rand, 
Stromberg Carlson, Sylvania Electric Products, Telemeter 
Magnetics, Texas Instrument, Westinghouse . . . 



2 * Computer BUYER Subscribers include: 



Arthur Anderson & Company, Blue Cross, Ebasco 
Services, Price Waterhouse, Bank of America, First 
National Bank of Boston, Bank of New York, Chase- 
Manhattan, First National Bank of Atlanta, Provident 
Trust, Equitable Life Assurance Society, Great Western, 
Metropolitan Life, Penn. Mutual, U.S. Life, Battelle 
Memorial Institute, Brookhaven National Laboratory, 
Bureau of the Census, National Bureau of Standards, 
Los Alamos Scientific Laboratory, National Weather 
Record Center, Port of New York Authority, Railroad 
Retirement Board, Rand, U.S. Air Force, U.S. Security 
Service, U.S. Army Tank Ordnance Center, U.S. Navy, 
David Taylor Model Basin, University of California, 
Kansas State College, Mass. Institute of Technology, 
San Diego State College, Wayne State University, 
University of Wisconsin, Ford Motor, General Motors, 
General Tire and Rubber, American Cyanamid, Eastman 
Kodak, Owens Corning Fiberglass, Monsanto Chemical, 
Allis Chalmers, Cutler Hammer, Diehl Mfg., Western 
Electric, Westinghouse, Calloway Mills, Kellogg, Briggs 
and Stratton, United Shoe, Atlantic Refining, Esso 
Research, Pan American. Sun Oil, Texas Eastern 
Transmission, Bethlehem Steel, Continental Can, U.S. 
Steel, American Tel. and Tel., Arizona Public Service, 
Bell Telephone Labs., Consolidated Edison, Florida 
Power, Pacific Gas and Electric, Johns Mansville, 
Boeing Airplane, Convair, Douglas Aircraft, Grumman, 
Hughes Aircraft, Lockheed, North American Aviation, 
Reaction Motors, American Airlines, Hawaiian Airlines, 
Pennsylvania Railroad. 



! 




Companies and Products Advertised in 

COMPUTERS and AUTOMATION 

(Below is a partial list ranging from 
September 1958 to September 1959) 



Advertiser 


Product Advertised 


Burroughs 


Computers, Components, Photoreader, 
Employment Opportunities 


RCA 


Transistors, Cores, 
Employment Opportunities 


ESC 


Pulse Transformers, Delay Lines 


Electronic Associates 


Computer, Variplotter 


General Electric 


Tubes, Capacitors 


Aircraft Nuclear Propulsion Dept. 


Employment Opportunities 


Computer Dept. 


Computers, Components 


Computer Capacitor Sales 


Capacitors 


Flight Propulsion Lab Dept. 


Consulting Service, 
Employment Opportunities 


Huntsville Computer Center 


Employment Opportunities 


Advanced Electronic Center 


Courses 


Apparatus Sales Office 


Power Supplies 


Heavy Military Electronics Dept. 


Employment Opportunities 


Ampex 


Tape, Tape Handlers 


C. P. Clare 


Relays, Wiring Services, 
Stepping Switches, Insulation 


Bendix Aviation 


Paper Tape Reader, Computers, 
Couplers, Employment Opportunities 


Ferroxcube Corp. 


Memory, Ferrite Cores 


Honeywell Datamatic 


Computers, Employment Opportunities 


Aeronutronic Systems 


Employment Opportunities 


Royal McBee 


Computers 


Ramo-Wooldridge 


Capacitors, Employment Opportunities, 
Computers, Computer Services 


System Development Corp. 


Employment Opportunities 


Arnold Engineering 


Cores, Magnets 


Audio Devices 


Tape 


Bryant Chucking Grinder 


Storage Drums 


Autonetics 


Employment Opportunities 


Southwest Research Institute 


Employment Opportunities 


Thompson-Ramo-Wooldridge Products 


Computers, Employment Opportunities 


Philco 


Computers 


Datics Corp. 


Computers 


Sylvania Electric 


Employment Opportunities 


Lockheed Aircraft 


Employment Opportunities 


Marquardt Aircraft 


Employment Opportunities 


Technical Operations 


Employment Opportunities 


Space Technology Lobs 


Employment Opportunities 


Harvey-Wells 


Building Blocks 


Corp. for Economic and Industrial 


Employment Opportunities, 


Research 


Consulting Services 


Di/An Controls 


Cores, Buffers 


NJE Corp. 


Power Supplies 


Computer Control Co. 


Memory, Modules 


International Business Machines 


Employment Opportunities 


Boeing Aircraft 


Employment Opportunities 


Reeves Soundcraft 


Tape, Kit to Check Recording Equipment 


National Cash Register Co. 


Computers, Circuitry, Memory 


Dialight 


Indicator Lights 


Whitnon 


Drums, Discs 


Raytheon Mfg. Co. 


Transistors 


Potter Instrument 


Computers and Data Processing Systems 


John Wiley & Sons 


Books 


Underwood 


Servotypers 


John Diebold 


Consulting Services 


Chrysler Corp. 


Employment Opportunities 


Republic Aviation 


Employment Opportunities 


Aerojet General 


Employment Opportunities 




COMPUTERS and AUTOMATION 

Will Be Read By Even More People Who Count 

During 1959, we plan that the circulation of COMPUTERS 
and AUTOMATION will reach at least 5000 copies, as 
compared with 2600 subscribers in January 1957, 3700 in 
January 1958, and 4000 for the June 1959 Directory Issue. 
A continuing circulation campaign promises to add large 
numbers of new subscribers to our audience. All new 
subscriptions are regularly paid for at $5.50 a year, assuring 
a top-quality as well as a top-quantity computer-minded 
audience. In addition, about 1000 key decision-makers 
working at computer builders and at computer buyers have 
been added on a controlled basis in order to assure 
circulation in depth. 




COMPUTERS and AUTOMATION 

Editorial Coverage is Expanding, Too. 

Each issue of COMPUTERS and AUTOMATION- 

includes new articles, new reference information, and new 
ideas, to make it as complete, informative, and stimulating 
to computer people as we can make it. Greater emphasis 
than ever before is being placed on news of computers and 
data automation . . . new surveys, new applications, new 
products, new devices, new ideas about computers and data 
processors . . . everything we can think of to help our readers 
do their jobs better, more quickly, and more profitably. 





I I I 



I I I I I 



I I I 



I I I I I I I I I I I I I I I I I 

COMPUTERS and AUTOMATION 



ADVERTISING REPRESENTATIVES 

Middle Atlantic States MILTON L. KAYE San Francisco 5 

535 Fifth Ave. New York 17, N.Y. 605 Market St. 

MUrray Hill 2-4194 

Washington 6, D.C. ROBERT CADEL Los Angeles 5 

1519 Connecticut Ave. Columbia 5-9727 439 S. Western Ave. 

Elsewhere THE PUBLISHER 

Berkeley Enterprises, Inc. 

815 Washington St., Newtonville 60, Mass. 

DEcatur 2-5453 or 2-3928 



A. S. BABCOCK 
YUkon 2-3954 

W. F. GREEN 
DUnkirk 7-8135 



[Continued from page 22] 

Special routines have been devised to deal with ho- 
monymous and polysemantic words; the analysis of the 
former words precedes that of the latter. 

The four principal types of "Homonyms” analyzed by 
the routine are those of "adjective-noun” (Homonym 1), 
"noun-verb” (Homonym 2), "verb-adjective” (Ho- 
monym 3) and of "preposition-adverb” (Homonym 4). 
Also, a more complicated sub-type is distinguished, that 
of "adjective-noun-verb,” as in "check,” which may be 
an adjective, a noun, or a verb. 

In specifying Homonyms 1, 2 and 3, a combination of 
a morphological and a syntactical analysis of the word is 
used. Thus, any inflection (except for ER or EST) iden- 
tified in Homonyms 1 or 3 makes "adjective” an im- 
possible alternative, just as ED or ING inflection in Ho- 
monym 2 deletes the "noun” possibility. These are 
morphological criteria, which do not, however, find as 
wide an application as syntactical analysis does in view 
of the scarcity of inflections in English. 

The information "Homonym” acquired in the course 
of this analysis may or may not be final for the diction- 
ary cycle, since some of the words in the MT dictionary 
are provided here with the indication "polysemantic” 
instead of a lexical equivalent. 

The total number of polysemantic words stored in our 
dictionary amounts to 500 words. 

The determination among multiple meanings of the 
single meaning of a word is accomplished by specifying 
typical contexts of polysemantic words in accordance 
with a special routine. This routine is the final one in 
the Dictionary Analysis of the English sentence. 

Grammatical Synthesis 

Grammatical processing of a sentence is broken up 
into two independent steps: Synthesis and Analysis. The 
former is the simpler. 

Synthesis routines provide rules for grammatical modi- 
fication of the translated text in accordance with gram- 
matical information obtained in the course of the analy- 
sis of the English original. 

The most important peculiarity of Synthesis routines 
is their "non-comparative nature.” This means that rules 
of word-changing, as well as certain rules of word-build- 
ing, are formulated strictly within a particular target 
language. Because of this, the same Synthesis routines 
can be applied to sentences translated from different 
languages. 

However, the requirements of synthesis routines are 
inclined to increase when they serve multi-lingual MT 
purposes. In such case, with multi-lingual MT in view, 
Synthesis routines should be: 

(1) EXHAUSTIVE, in describing the word-changing 
system in the target language, since grammatical 
rules with no application in MT from one language 
may become vitally important when the source lan- 
guage is changed; 

(2) INFALLIBLE, so as to carry through any instruc- 
tion obtained from the Analysis of the source lan- 
guage. This requires providing every "non-pro- 
ductive” category of the target language with a 
"productive” grammatical equivalent. 

So far, the problem of grammatical equivalents 
within a language, theoretically, stands out as the 
most important for Synthesis. 

COMPUTERS and AUTOMATION for October, 1959 



(3) INDEPENDENT of the Analysis, since the iat- 
ter may be very different for different languages. 

Grammatical Analysis 

Unlike Synthesis, independent Analysis cannot be 
recommended, for this would not help at all to make 
MT economical. Analysis problems are numerous, and 
are important scientifically. We shall discuss one of them 
at some length. 

English Analysis is covered by six routines, of which 
we shall give a general outline of one, "Verb.” It stands 
out as a routine playing a key part in the whole pro- 
cedure of Analysis. 

The "Verb Analysis” Routine 

The ‘Verb Analysis’ routine is divided into five sec- 
tions; the first section is compulsory for every verb of 
the sentence, whereas of the remaining four sections, 
only one is used for each type of the analyzed verbs. 

In Section I, verb selection for further analysis is per- 
formed. Among words picked out for analysis in this 
routine are those possessing the indication "Verb,” so 
far as they do not have any of the following indications: 
"to be Disregarded,” "Not to be Changed,” or (Rus- 
sian) indications "Participle,” "Verbal Adverb” or 
"Verbal Noun.” The checking for absence of these in- 
dications is meant to exclude from further analysis those 
verbs that have been elsewhere provided with charac- 
teristics that satisfy the Synthesis routines. 

In addition to verb selection, correction of certain verb 
indications is provided for in Section I. 

Among verb indications liable to correction are those 
of tense with verb-predicates in if-clauses, and of case 
government with link-verbs, as well as some more par- 
ticular indications. Analysis of homonymous forms, 
such as Past Indefinite and Subjunctive, and of irregular 
verbs also belongs here. 

Checking for grammatical context, implying correc- 
tion as a possibility, is also performed, both when one 
of the above-mentioned indications is ascribed to the 
analyzed verb in the dictionary and also when it is about 
to be developed in the course of further Analysis. 

The preliminary checking of Section I is followed by 
verb analysis proper. For this purpose the analyzed verb 
is sent to one of the four different sections, differences 
in morphological structure of the verb being decisive in 
choosing the section. Thus, verbs with S-ending (affix S 
following the stem of the word) are sent to Section II ; 
verbs with ED-ending, as well as certain forms of ir- 
regular verbs, enter Section III; and verbs with ING- 
ending are directed to Section IV ; verbs not inflected are 
analyzed in Section V. 

S-Verbs 

Grammatical qualification of 'S-verbs’ in Section II 
depends on whether S stands out as the only ending of 
the verb, or another ending (usually ING) is associated 
with it. In the latter case, the following indications are 
developed for the Russian equivalent verb: "Verbal 

Noun; Neuter; Plural,” which implies further analysis 
by the "Noun” routine at the proper time. 

When S is the only ending, English characteristics of 
the verb (Predicate in the Present Indefinite form) are 
transformed into Russian indications, but not without 
checking for correction conditions. Resultant character- 
istics are "Predicate,” associated with either "Present” or 



27 




"Future tense/’ "Number” and "Person” (or gender for 
the Past tense in other cases). But some characteristics 
of the Russian predicate remain not defined until the 
subject of the Russian sentence is determined. 

-ED-Verbs 

The analysis of "ED-verbs,” i.e., verbs with ED-ending 
and certain groups of irregular verbs, is performed in 
Section III. Here syntax definitely takes precedence, and 
four main patterns of grammatical verb context are here 
analyzed. 

Noteworthy is the fact that context analysis of a word 
implies, in all cases, observation of "Rules of Word 
Selection.” These rules are based on classifying all the 
words in a sentence into three categories, which are: 

(1) words of third-degree structural significance; parti- 
cles, adverbials, parentheses, and parts of a sentence 
combined by a coordinating conjunction, etc., are 
included ; 

(2) words of second-degree structural significance, 
where different words and word groups belong, so 
far as they are placed in the attributive position to- 
wards some word of a sentence; 

(3) words of first-degree structural significance, which 
include words not identified as belonging to either 
of the two previous categories. 

Through application of "Rules of Word Selection” in 
the course of the searching procedure all words of lesser 
category than the word searched are omitted; the chief 
constituents of the grammatical pattern required are 
thus singled out. 

We cannot here give a detailed description of all the 
processes involved in the analysis of verb patterns in Sec- 
tion III, but we can comment on a few of the patterns 
that lead to interesting solutions. 

For example, in transforming English construction of 
Modal Passive: 

Modal Verb plus Selected Verb indicated, and "Auxili- 
ary I” (BE) plus Analyzed Verb, into Russian Active 
Compound predicate : 

Modal Verb; Impersonal plus Analyzed Verb, In- 
finitive, the transformation is associated with conver- 
sion of English subject into Russian Direct Object. 

Also, another pattern is transformation of English 
Complex Object Construction into Russian subordinate 
clause. 

Resultant characteristics developed for the verbs ana- 
lyzed in Section III include both morphological and syn- 
tactical information. Of syntactical indications only 
"Predicate” and "Attribute” are fixed here. The former 
is associated with morphological indications of mood 
(Indicative, Subjunctive and Infinitive are developed 
here), of tense (Present or Past), and of voice (both Ac- 
tive and Passive) . 

The indication "Attribute” is accompanied by mor- 
phological indications of "Participle,” tense (Present or 
Past), and voice (Active or Passive). 

-ING-Verbs 

ING-forms of verbs are defined in Section IV. Here 
the same verb patterns and others besides are analyzed, 
though important changes in their value affect the order 
in which they are searched here. 

The resultant characteristics of the Russian equivalent 

28 



verb include one of the following sets of indications: 
(1) "Verbal Noun, Neuter”; (2) "Participle, Present 
tense. Active voice; Attribute”; (3) "Verbal Adverb, 
Present (or Past) tense”; (4) "Not to be Translated,” 
"to be Disregarded.” In addition to these, "Infinitive,” 
"Subjunctive” or "Indicative mood,” with the corres- 
ponding set of indications, are examined here in case 
the analyzed verb takes its characteristics from some of 
the "selected helping words.” 

Verbs Not Inflected 

Verbs not inflected are analyzed in Section V, with 
various verb patterns. One pattern which should be 
specially mentioned deals with transformation of the 
English constructions of Complex Subject and of at- 
tributive Infinitive into the Russian complex sentence 
or subordinate clause, accordingly. 

The resultant information here includes the indica- 
tion of Infinitive, Imperative, Subjunctive, or Indicative 
mood, with the indications of tense (Present, Past or 
Future) and voice (Active or Passive) attached in case 
of the Indicative mood. The only syntactical indication 
fixed here is "Predicate.” 

Another section of the verb analysis routine is based 
on a classification of verbs, devised to characterize Eng- 
lish verbs both within the English system and with re- 
gard to Russian translating conventions. 

Within the English language verbs are classified into 
MODAL and HALF-MODAL (help, dare), AUXILI- 
ARY and 7 sub-classes of HALF- AUXILIARIES, CAUS- 
ATIVE (cause, enable, make, order, command, etc.), 
DECLARATIVE (declare, call, label, report, etc.), Verbs 
Taking two Objects (give, offer, permit, etc.), etc. 

To meet the requirements of the Russian translating 
conventions, verbs are divided into classes and semantic 
groups. To date, 53 groups of verbs have been estab- 
lished. These are summarized into three classes, the first 
two classes comprising verbs having translational pecu- 
liarities in Finite (class I) or Infinite (class II) forms; 
class III covers more complicated cases. 

The "Verb Analysis” routine is applied until every 
verb of the sentence is provided with all the grammatical 
information required in the Synthesis routines, except 
for the indications of number, person (or gender), which 
are not defined until the subject of the Russian sentence 
is established. 

Noteworthy is the fact that the information obtained 
in this routine is not restricted to the analyzed verb, but 
is extended to cover the information available at this 
stage of Analysis, concerning "selected” (helping) 
words (verb, nouns, adjectives, etc.) and punctuation 
marks. 

Moreover, quite a number of transformations in sen- 
tence structure are introduced here. These include 
change of word-order, inserting necessary conjunctions 
and other words or punctuation marks, etc. There are 
transformations associated with the translation of Com- 
plex Subject and Complex Object, Attributive Infinitive 
and Gerundival Subject, as well as some other verb con- 
structions. 

And, of course, there is much more to be said besides. 
But perhaps enough has been said to show some of the 
main lines of our current research, thinking and aims 
in the subject of machine translation of languages. 

COMPUTERS and AUTOMATION for October, 1959 




Make over 200 Small 
Computing and Reasoning 
Machines with . . . 



AIN I AC 



NEW 1959-60 ELECTRIC BRAIN CONSTRUCTION KIT 



WHAT COMES WITH YOUR BRAINIAC KIT? All 33 experiments from our original Geniac kit (1955), with 
exact wiring templates for each one. All 13 experiments from the former Tyniac kit, with exact wiring templates. 
156 entirely new experiments (Brainiacs) with their solutions. Over 600 parts, as follows: 6 Multiple Switch Discs; 
Mounting Panel; 10 Flashlight Bulbs; 2 Multiple Socket Parts, each holding 5 bulbs; 116 Wipers, for making good 
electrical contact (novel design, patented); 70 Jumpers, for transfer contacts; 50 feet of Insulated Wire; Flashlight 
Battery; Battery Clamp and Main On-Off Switch; nuts, bolts, sponge rubber washers, hard washers, screwdriver, spin- 
tite blade, etc. ALSO: an introduction to Boolean Algebra for designing circuits; booklet "How to go from Brain- 
iacs and Geniacs to Automatic Computers”; complete descriptions of 201 experiments and machines; over 170 cir- 
cuit diagrams; list of references to computer literature. 

This kit is an up-to-the-minute introduction to the design of arithmetical, logical, reasoning, computing, puzzle-solving, 
and game-playing circuits — for boys, students, schools, colleges, designers. It is simple enough for intelligent boys to as- 
semble, and yet it is instructive even to engineers, because it shows how many kinds of computing and reasoning circuits 
can be made from simple components. This kit is the outcome of 10 years of design and development work with Geniacs 
and small robots by Berkeley Enterprises, Inc. With this kit and manual you can easily make over 200 small electric brain 
machines that display intelligent behavior and teach understanding first-hand. Each one runs on one flashlight battery; all 
connections with nuts and bolts; no soldering required. (Returnable for full refund if not satisfactory.) . . Price $18.95 



WHAT CAN YOU MAKE WITH A BRAINIAC KIT? 

LOGIC MACHINES 
Syllogism Prover 
James McCarty’s Logic Machine 

AND, OR, NOT, OR ELSE, IE . . . THEN, IF AND ONLY 
IF, NEITHER . . . NOR Machines 
A Simple Kalin-Burkhart Logical Truth Calculator 
The Magazine Editor’s Argument 
The Rule About Semicolons and Commas 
The Farnsworth Car Pool 

GAME-PLAYING MACHINES 
Tit-Tat-Toe 
Black Match 
Nim 

Sundorra 21 

Frank McChesney’s Wheeled Bandit 

COMPUTERS — to add, subtract, multiply, divide, . . . . , 
using decimal or binary numbers. 

— to convert from decimal to other scales of notation and 
vice versa, etc. 

Operating with Infinity 

Adding Indefinite Quantities 

Factoring Any Number from 4 5 to 60 

Prime Number Indicator for Numbers 1 to 100 

Thirty Days Hath September 

Three Day Weekend for Christmas 

Calendar Good for Forty Years 1950 to 1989 

Money Changing Machine 

Four by Four Magic Square 

Character of Roots of a Quadratic 

Ten Basic Formulas of Integration 

PUZZLE-SOLVING MACHINES 
The Missionaries and the Cannibals 
The Daisy Petal Machine 
Calvin’s Eenie Meenie Minie Moe Machine 
The Cider Pouring Problem 
The Mysterious Multiples of 76923, of 369, etc. 

Bruce Campbell’s Will 

The Fox, Hen, Corn, and Hired Man 

The Uranium Shipment and the Space Pirates 

General Alarm at the Fortress of Dreadeerie 

COMPUTERS and AUTOMATION for October, 1959 



The Two Suspicious Husbands at Great North Bay 

The Submarine Rescue Chamber Squalux 

The Three Monkeys Who Spurned Evil 

Signals on the Mango Blossom Special 

The Automatic Elevator in Hoboken 

Timothy’s Mink Traps 

Josephine’s Man Trap 

Douglas Macdonald’s Will 

Word Puzzle with TRICK 

QUIZ MACHINES 

The Waxing and the Waning Moon 
Intelligence Test 
Guessing Helen’s Age 
Geography Quiz 
Mr. Hardstone’s Grammar Test 
Solving Right Triangles 

SIGNALING MACHINES 
The Jiminy Soap Advertising Sign 
The Sign that Spells Alice 

Tom, Dick, and Harry’s Private Signaling Channels 
Jim’s and Ed’s Intercom 

CRYPTOGRAPHIC MACHINES 
Secret Coder 
Secret Decoder 

Lock with 65,000 Combinations 
Lock with 15,000,000 Combinations 
The General Combination Lock 
Leonard’s Two-Way Coding Machine 

. . . AND MANY MORE 

I . . - _ Mail this Request or a copy of it 

| Berkeley Enterprises, Inc. 

I 815 Washington Street, Rl43, Newtonville 60, Mass. 

, Please send me BRAINIAC KIT K18, including manual, 
instructions, over 600 parts, templates, circuit diagrams, etc. 

, I enclose $18.95 for the kit plus for handling and 

I shipping (30c, east of Mississippi; 80c, west of Mississippi; 

I $1.80, outside U.S.). I understand the kit is returnable in 

I seven days for full refund if not satisfactory (if in good 

I condition). 

1 My name and address are attached. 



29 




Readers 9 and Editor’s Forum 

[Continued from page 6] 

2. What are your suggestions for the name of this 

group ? 

3. In what ways do you think computer people can be 
of particular help to the rest of society in influencing 
for good the applications and effects of computers? 



4. In what ways would you like to work? 

5. Any remarks? 

Name 

Address 

When you have filled this in, please send it to E. C. 
Berkeley, Acting Secretary, WGSDAC, 815 Washington 
St., Newtonville 60, Mass. 

ON FREEDOM TO PUBLISH 

I. From: James A. Cook, The Rand Corporation, 
Santa Monica, Calif., to the Editor. 

The RAND Corporation has always been interested 
in the widest possible dissemination of those research 
studies which are of value to the public. This purpose 
was being served, we felt, upon seeing your July issue 
with its report by Allen Newell and J. C. Shaw of 
RAND and H. A. Simon of Carnegie Tech, entitled 
"A General Problem-Solving Program for a Computer.” 
All of us, including the authors, felt it was a genuine 
compliment that you had seen fit to include this paper 
among the three selected for publication. The au- 
thors expressed dismay, however, that the title used by 
Computers and Automation gave an appearance of 
finality to their work that they did not wish to imply, 
and they had selected a title commencing "A Report 
On . . .” because of the interim nature of their research. 

A further discussion of the report revealed that it 
seemed to have been reproduced in your magazine with- 
out consultation or approval by the authors. We have 
not sought to restrict the report from the public domain. 
But, we do question the wisdom, both from the stand- 
point of reflecting up-to-date research and of business 
courtesy, for a commercial magazine to publish in toto 
a scientific paper without a prior inquiry or notification 
of the authors. 

If you had an assent for publication from a source 
of which I’m unaware, my apologies for the assumption. 

II. From the Editor to Mr. Cook: 

Thank you for your letter of August 11. I am sorry 
for the misunderstanding about the paper by Newell, 
Shaw, and Simon. We shall of course publish a correc- 
tion that deals with the meaning of "A Report On . . .”, 
and we shall be glad to publish any other corrections 
that you and they feel should be made — may I please 
hear from you quickly as to any other changes? 

In regard to your other point, I do sympathize with 
the feelings of authors who find themselves suddenly in 
print when they did not intend to be. But in the op- 
eration of a magazine a great deal of information flows 
across the desk of an editor; often, with pressure of 
deadlines upon him, he has to make quick decisions 
about what is the most interesting material that he has 

,30 



to print. Whenever a piece is marked "confidential” 
or "not to be printed” or "copyright”, no reputable 
editor does print it, without securing permission. There 
is of course much other literature which the senders 
thereof are very eager to have printed. 

In the case of the papers for the International Con- 
ference on Information Processing, there was nothing at 
any time in any of the literature that indicated the 
papers were not in the public domain. In fact, I do not 
believe that UNESCO would have consented to accept 
any paper for presentation that could not have been 
freely released to the press at the time of the conference. 
All of the 1700 people in Paris attending the conference 
were given an armful of preprints, and so were the 
press representatives; and I am positive that there will 
be printing or reprinting of parts or all of at least some 
of the papers, in many countries, Russia, France, Eng- 
land, etc. 

1 regret very much that apparently Mr. Newell, Mr. 
Shaw, and Mr. Simon did not realize the circumstances 
under which their paper was being released to the press. 

If any similar occasion comes up in the future in- 
volving these authors, we shall now of course inquire 
to see if they mind being reprinted. However, this will 
place us at a disadvantage in regard to other organiza- 
tions of the press, who in the absence of legal notice or 
some equivalent, will feel that they can freely quote or 
reprint. There is a legal implication that a person who 
does not take the trouble to put the word "copyright” 
on what he is writing or an equivalent notice, thereby 
does not restrict the copying of the information. 

With best wishes . . . 

III. Telegram from Mr. Cook to the Editor: 

Appreciate your 17 August letter. The one correction 
you are publishing will suffice. Regards. 

I960 WESTERN JOINT COMPUTER CONFERENCE 
— INVITATION FOR PAPERS 
H. M. Zeidler 

Technical Program Committee Chairman, I960 WJCC 
Stanford Research Inst. 

Menlo Park, Calif. 

"Computers — Challenge of the Next Decade” is the 
theme of the I960 Western Joint Computer Conference 
scheduled for next May 3, 4, 5 in San Francisco, 
California. 

Special emphasis will be placed on major areas where 
new planning, and new research and development pro- 
grams are to be directed for the computer growth 
destined for 1960-70. For example, papers on concepts 
and techniques in newer areas such as language trans- 
lation, data retrieval, and self-teaching systems would 
be of particular significance. It is of course intended 
that the conference shall also be a base for technical 
papers and discussions pertaining to the current state 
of the computer art — both analog and digital. 

The Conference Proceedings incorporating the papers 
to be presented at the I960 WJCC will be distributed 
at the registration desk. They will be of value to the 
conference attendee for his post-conference study, but 
also for his current comprehension of the technical 
sessions. Papers to be submitted to the Technical Pro- 
gram Committee should be prepared for thirty minutes 
delivery, extra time being available for discussion. Eval- 

COMPUTERS and AUTOMATION for October, 1959 




uation will be based on study of a first draft of the 
complete paper, three copies of which should be in the 
hands of the program committee as early as possible, and 
by 9 November, 1959, at the latest. No advance abstract 
of the paper will be required, but notification of intent 
to submit a paper should be provided as soon as pos- 
sible to facilitate program planning. The papers will 
be reviewed, final selection made, and authors notified 
by 25 January, I960. The papers must then be received 
in final form by the Technical Program Committee by 
1 March for incorporation into the Conference Pro- 
ceedings. 

NOTICE FROM A COMMITTEE ON 
DATA SYSTEMS LANGUAGES 

A. E. Smith 

Chairman, Intermediate-Range Task Force 
Committee on Data Systems Languages 
Bureau of Ships, Department of the Navy, 
Washington 25, D.C. 

Would you please publish the attached notice in the 
next issue of Computers and Automation? 

The Intermediate-Range Task Force of the Committee 
on Data Systems Languages would like to hear from 
people who wish to participate in the work of the Task 
Force as associate members. The purpose of associate 
membership is to provide the Task Force with the bene- 
fit of the experience and thinking of people engaged 
in all aspects of business data processing applications. 
The particular mission of the 1RTF is to specify a 
common language for describing data processing sys- 
tems. Associate members will not normally be asked to 
participate extensively in the working efforts of the 
Task Force or its task groups, but will be recipients of 
the publications of the Task Force and will, from time 
to time, be called upon to respond to surveys, give 



opinions, or otherwise guide the thinking of the work- 
ing force. 

Persons interested in becoming associate members 
are asked to write to Mr. A. E. (Gene) Smith, Chair- 
man, Intermediate-Range Task Force, Code 280, Navy 
Department, Bureau of Ships, Washington 25, D.C., 
giving a statement of interest and experience. 



MATHEMATICIAN 

with graduate study and computer (650) ex- 
perience for opportunity position. Applicants 
must possess ability to utilize literature on 
numerical methods in the computer solution of 
engineering problems. 

Matrix algebra, linear differential equations, and 
statistical analysis involved. 



Chrysler Corporation 
Engineering Division 
P. O. Box 1118 
Detroit 31, Michigan 



EARLY ISSUES OF 
"COMPUTERS AND AUTOMATION” 

I. From Dr. Herbert Rister 
Bibliotheksrat 
Westdeutsche Bibliothek 
Marburg, Germany 

The Editorial Staff of the "Gesamtverzeichnis der 
Auslandischen Zeitschriften und Serien,” which has been 
processed at the Westdeutsche Bibliothek, Marburg/ 



Lahn, Germany, would greatly appreciate your letting us 
know dates of change of title of Computers and Auto- 
mation which formerly had the title "Computing Mach- 
inery Field.” Please inform us, besides, about number- 
ing of volumes and exact date of change of title (day, 
month, year). 

II. From the Editor 

The early issues of Computers and Automation were 
as follows: 



Designation 




Title 


Date 


No. of 
Pages 


Manner of 
Publication 


Vol. 1, no. 


1 


Organizations in the Field of Automatic Computing 


Aug. 15, 1951 


7 


Purple Ditto 


Vol. 1, no. 


2 


Machinery 

Roster of Organizations in the Field of Automatic 


March, 1952 


8 


Purple Ditto 


Vol. 1, no. 


3 


Computing Machinery 

Roster of Organizations in the Field of Automatic 


July 20, 1952 


11 


Purple Ditto 


Vol. 1, no. 


4 


Computing Machinery 
The Computing Machinery Field 


Oct., 1952 


36 


Photo-Offset 


Vol. 2, no. 


1 


The Computing Machinery Field 


Jan., 1953 


44 


Photo-Offset 


Vol. 2, no. 


2 


Computers and Automation (all later issues have had 


March, 1953 


36 


Photo-Offset 




this title) 









Beginning with 1953, all volumes of Computers and Automation coincide with the calendar year. Beginning with 
1955, there were at least 12 monthly issues in each year. In some years there have been extra issues. 



COMPUTERS and AUTOMATION for October, 1959 



31 





Air Traffic Control in the Year 1980 

Vernon I. Weihe 

Director of Air Traffic Control and Navigation Programs, 

General Precision Equipment Corp., 

Washington 5, D.C. 

(A talk given at The First World Congress of Flight, National Business 
Aircraft Association Symposium, Las Vegas, Nevada, April, 1959) 



I T is now April 1980, and although I have reached 
the allotted span — three score years and ten — it 
gives me great pleasure to discuss the present air traffic 
control system with you. As you well know, this is the 
Twenty-First World Congress of Flight, and each year 
since that first meeting, way back in 1959, has been a 
milestone in the onward march of aviation. 

Twenty years ago, many of us feared that a nuclear 
war might actually blot out the civilized world. If this 
had happened, the subject of this meeting might more 
appropriately concern the matter of "homesteading” 
the battered remnants of the world’s population on 
some uncontaminated planet of the solar system. 

Fortunately for all people the world over, the last 
two decades (since 1959) have been marked by a period 
of relative peace, and by an increasing awareness of the 
need for safer and more flexible use of aircraft. The 
Mach 3 transport airplane has become the basic me- 
dium for international person-to-person communication 
and, as such, is a major tool in the forging of better un- 
derstanding among the nations and the peoples of the 
world. 

Global Air Navigation and Traffic Control 

Independent national and regional planning and in- 
stallation programs in air navigation and traffic control 
have become largely outmoded; the renovated world or- 
ganization has reached a more mature professional sta- 
tus. Local, or regional faits accomplis are no longer de- 
bated by "rigidly briefed” delegates. Planning now 
starts at the system engineering level, with global as- 
pects well in mind. Experts from all interested nations 
meet in constructive conclave to pool their information 
and experience, and to jointly solve problems on a com- 
mon and compatible basis. Because of the completeness 
and soundness of the plans, national legislative groups 
the world over now provide a new level of continuity in 
leadership and program support. Feast and famine fin- 
ancing, aborted programs, and the recurring dispersal 
of expert air navigation and traffic control systems per- 
sonnel are no longer permitted by the top policy makers 
of the various administrations. 

During the past twenty years, notable progress has 
been made in two very important "slum clearance” pro- 
grams. The 1960-1970 period was marked by an effect- 
ive clearance of airspace slums, followed by the construc- 
tion of more satisfactory criteria and better methods of 
airspace use. The 1970-1980 decade was marked by a 
more complex, but no less important, "slum clearance” 
of the radio frequency spectrum. 

Airspace Slum Clearance 

The first program — that of "airspace slum clearance” 



Mr. Weihe is an authority in the field of air 
traffic control; many of his predictions on air 
traffic control in the past have been remarkably 
borne out by developments. His experience before 
joining General Precision included service as 
chief engineer of the Communications and Navi- 
gation Laboratory, Wright Air Development Cen- 
ter ; Air Traffic Control and navigation systems en- 
gineer for the Air Transport Association of Amer- 
ica; and U.S. delegate to the International Civil 
Aviation Organization. 



— was made possible by the formation of the Federal 
Aviation Agency. Its programs in the application of 
electronic digital computers provide a workable air- 
space inventory and file system. No longer are large 
areas of airspace withdrawn from common user status 
(and thereby markedly reduced in overall value) to suit 
an infrequent, or intermittent special purpose. The elec- 
tronic filing system contains an agreed priority system — 
when an airspace user completes his operation, the air- 
space automatically is made available to the next lower 
priority user who has filed a request or a flight plan. 

The airspace inventory and reservation system now al- 
lows a substantial reduction in longitudinal separation 
over the old ten-minute rule of the 1960’s. The "co- 
coon” of airspace around an aircraft is now related to 
the geographic locale and the amount of traffic involved. 
An aircraft no longer requires a billion times, or even a 
million times, its physical volume as an airspace "co- 
coon” for safe separation from other aircraft. Each air- 
craft is now more nearly able to fly the optimum path — 
safety, expeditiousness, and comfort are assured. Some 
aspects of this system will be described a bit later in 
this talk. 

Radio Spectrum Slum Clearance 

The second "slum clearance” program was completed 
during the last decade — 1970-1980. It concerns that 
basic resource which is the sine qui non of all subsystems 
which utilize transmitted radio energy. This resource 

— the radio frequency spectrum — is the real estate on 
which critically important aviation subsystems must be 
built. About thirty-five years ago, at the end of World 
War II, there was a short "land-rush” by many poten- 
tial user groups. Allocations were made and defended 
on the basis of "I asked for it first!” Little fundamental 
knowledge was applied to the subject; in fact, in most 
cases fundamental system design information was lacking 
and, more importantly, the impact on practical economic 

COMPUTERS and AUTOMATION for October, 1959 



32 





A\XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX\XXXXXXXXXXVVXXVXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX\YXXXXXXXXXX\yV 






MITRE 



C ORPORATIO N 



Incites 



COMPUTER PROGRAMMERS 

O o Join Its *Exf angina 
Scientific Community 



MITRE is a non-profit organization formed under the sponsorship of the 
Massachusetts Institute of Technology. Its nucleus is a Technical Staff 
of scientists and engineers with established reputations in the design and 
development of computer-based systems. 

The organization’s task is to provide the technical solutions to problems 
in the design, development and integration of large-scale, real-time control 
air defense systems. Ranging in scope from the practical problems of 
improving present systems through the design of future systems, MITRE 
assignments provide a unique multi-discipline environment in which 
significant contributions can be made. 

System-oriented programmers with experience on large-scale computers 
will find opportunities to define problems and formulate solutions in: 



• REAL TIME SIMULATION 

• SYSTEM PROGRAMMING 

• MATHEMATICAL MODEL MANIPULATION 

• COMPUTER APPLICATIONS 



• MACHINE DESIGN 

• INFORMATION STORAGE and RETRIEVAL 

• CODING TECHNIQUES 

• SUBSYSTEM SIMULATION 



• AUTOMATIC PROGRAMMING RESEARCH 

These positions are available at MITRE’s modern facilities in suburban 
Boston, Massachusetts, Fort Walton Beach, Florida, and Montgomery, Alabama. 



To arrange an immediate confidential interview, 
please direct your inquiry to: 

Dana N. Burdette, Personnel Director 

THE MITRE CORPORATION 

244 Wood Street — Lexington 73, Massachusetts 

A brochure more fully describing ‘MJJKE and its activities is available on request. 



H\xv 



NVXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX** 



COMPUTERS and AUTOMATION for October, 1959 



33 



A FULL LINE OF THE MOST 
RELIABLE, MOST DURABLE 
RELAYS EVER MADE 

P Nevw *•; # - £*£•**• 

weir Out' , ; 'or Boynter.^;” :i.;Ma8^efiaac«^^ 

»■ • - •* .* ■■■* ■.*„*** te * -s * » * ir> »^v* .*.%*•. 




single form D 

Type HG. Capable of up to 100 
operations a second. Load- 
handling capacity to 5 am- 
peres and up to 500 volts. (250 
va max. with proper contact 
protection.) 




single form D 

Type HGS. Biased with per- 
manent magnets. Speed up 
to 200 cps. Sensitivity as low 
as ±2.5 milliwatts. Handles 
up to 2 amperes, up to 500 
volts (100 va max. with proper 
contact protections.) 




single form D 

Type HGP. Can be factory ad- 
justed to provide single side- 
stable, bi-stable or chopper 
characteristics. 




two or three form D 

Types HG2 and HG3. Two or 
three form D Switches en- 
closed in a single housing. 
HG2 has 8 or 11-pin octal style 
plug. An 11-pin base is 
standard for HG3. 




Type HG4. Four form D 
switches enclosed in a single 
housing with coll. Plug in as- 
sembly (shown) is standard. 
Other mountings available. 




six in-line flat pack 

Type HG6F. Six switches 
mounted in line on printed 
circuit panel surrounded by 
single coil. Flat, compact as- 
sembly. Over-all dimensions: 
3.640" x 3.125" x 1.046”. Uses 
standard 32 or 36 terminal 
printed circuit socket. 




For complete information on clare 
Mercury- Wetted Contact Relays or 
on the entire clare line of superior 
relays and electronic components 
address: C. P. Clare & Co., 3101 
Pratt Blvd., Chicago 45, Illinois. 

In Canada: C. P. Clare Limited, 

P. O. Box 73, 2700 Jane Street, 
Downsview, Ontario. Cable Address: 

CLARELAY. 

Send for Catalog 




factors of service use were not understood. Wasteful 
systems were permitted to remain in operation, large 
areas were blocked off for uses which did not materialize, 
and selfish interests were allowed to hold deed to blocks 
of spectrum space which they were not occupying and 
stood little chance of requiring. This forced important 
systems to be located at awkward, compromise loca- 
tions where system optimization was not feasible. 

As recently as twenty years ago, there were no channel 
allocations, on a user basis, in the radar bands. Obsolete, 
or obsolescent, low stability systems were allowed to 
"wild cat” or deteriorate newer, high-performance, high- 
stability systems. Some groups argued for a particular 
spectrum allocation because it was free of weather ob- 
scuration and good for "seeing” aircraft, while others 
argued for the same spot in the spectrum for "seeing” 
weather and not aircraft. Likewise, in navigation, cer- 
tain air/ground cooperating systems occupied bands too 
high for the service because obsolete or obsolescent sys- 
tems of earlier vintage were still in limited use and no 
administrative machinery existed for clearing them out. 

The radio spectrum "slum clearance” program in the 
United States was brought about in a manner quite 
similar to that of the airspace "Slum Clearance” pro- 
gram. The Federal Aviation Agency legislation which 
abolished the separate control of airspace by Civil and 
Military authorities, formed a model for the legislation 
which attacked the radio spectrum problem. The prob- 
lem was quite a bit more difficult, however, since the 
radio spectrum problem was by no means confined to 
groups having a common interest, such as aeronautics. 
In any event, through executive and legislative leader- 
ship, in 1966, a single agency was formed to replace the 
two who formerly had cognizance. The new agency was 
called the Federal Radio-Emission and Communications 
Agency (FRACA), and was given "power of eminent 
domain.” Funds were made available to buy up obsolete 
or obsolescent equipments and additional means were 
found to aid in the development, production, and im- 
plementation of new equipments and systems. Time 
schedules were set and they were met. Recalcitrant users 
of outmoded equipments were ordered "off the air.” 

Radars Where They Belong 

The results we see today — weather radars are up 
where they belong, aircraft surveillance radars are down 
where they belong. Standard LORAN, and four-course 
ranges have been cleared out, and a new system having 
a nearly optimum radio frequency allocation has been 
implemented for both land and sea area coverage. 

The single authority has brought new impetus to non- 
government programs. For ten years now, the single 
authority, working through its overseas counterparts, 
has left no doubt concerning the integrity of world-wide 
radio frequency service allocations. In the common car- 
rier radio telephone service, because of the decreased 
risk, large amounts of development and implementa- 
tion capital have been made available. We now — in 1980 
— have an optimally located air/ground common car- 
rier radio telephone service. Today, any occupant of an 
aircraft aloft on any of the world’s airlanes, can dial to 
any other subscriber on land, sea, or aloft. The system 
is essentially world-wide. The high quality solid state 
trans-oceanic radio telephone cables and the wide-band, 
non-orbiting satellite communication relay stations vie 

COMPUTERS and AUTOMATION for October, 1959 



34 





...another example of unequaled “in-process” controls for 



RCA MEMORY CORES 



Unusually stable temperature during “firing” of RCA Cores makes possible 
extremely high uniformity of electrical characteristics . . . provides you with 
RCA Cores to meet your most rigid specifications / 



The electrical characteristics of a 
memory core are directly deter- 
mined by the temperature at 
which the ferrite material is fired. 
The more constant the tempera- 
ture, the more uniform the elec- 
trical characteristics. And the 
electrical furnaces in use at 
RCA’s Needham Heights ma- 
terials lab are controlled to within 
± 0.5°C at temperatures as high 
as 1350°C. This precise control 
means not only high uniformity 
from core to core. . .but the tightest 
production controls on your cores 



from order to order. 

This extremely efficient firing 
technique is still another example 
of “in-process” controls employed 
by RCA in every stage of manu- 
facture from the selection of raw 
materials to the inspection of fin- 
ished product. For Ferrite Cores, 
Planes, and Stacks to meet your 
specific design requirements, con- 
tact your local RCA Field Repre- 
sentative. For technical data, 
write RCA Commercial Engi- 
neering, Section J-90-NN , Somer- 
ville, New Jersey. 




Unique memory-plane and stack tester 

dynamically tests every RCA core in every RCA 
plane while the entire stack is being driven 
under actual use conditions. ..provides 100% 
testing at actual computer frequencies. 

East: 744 Broad Street, Newark, N. ]. 

HUmboldt 5-3900 
Northeast: 64 ‘‘A” Street, Needham Heights 94, Mass. 
H I Merest 4-7200 

East Central: 714 New Center Bldg., Detroit 2, Mich. 
TRinity 5-5600 

Central: Suite 1 154, Merchandise Mart Plaia 
Chicago, III. WHitehall 4-2900 
West: 6355 E. Washington Blvd., 

Los Angeles, Calif. RAymond 3-8361 
Gov't: 224 N. Wilkinson St., Dayton, Ohio 
BAIdwin 6-2366 

1625 "K” Street, N.W., Washington, D.C. 
District 7-1260 



RADIO CORPORATION OF AMERICA 

SEMICONDUCTOR AND MATERIALS DIVISION • SOMERVILLE, N. J. 



COMPUTERS and AUTOMATION for October, 1959 



35 




with each other to determine which can provide the most 
reliable and clearest point-to-point telephone service. 

In the foregoing, I have alluded to changes in leader- 
ship, and to the two "slum clearance’’ programs which 
have been accomplished since the time when the First 
World Congress of Flight was held, back in 1959- I 
shall now summarize some of the operational and tech- 
nical advances since that time. 

Revolution in Electronic Computers 

The fifteen-year period following World War II 
(1945 to i960) was marked by a revolution in the elec- 
tronic digital computer field. During the first portion, 
the computer engineer was adapting radio and radar 
components and circuits for computer uses, but toward 
the end of the period, under the impact of advances in 
solid state physics, unique components and circuits es- 
pecially applicable to the computer art began to appear. 
At first, transistors and diodes replaced radio tubes, and 
printed circuit boards replaced wired component sub- 
assemblies. Then there was an attack on storage or mem- 
ory subsystems. The subsystem became smaller, faster, 
and more flexible for any given size task. This was fol- 
lowed in the 1960-1980 period by advances in micro- 
miniaturization, based on breakthroughs in field emission 
and thin film techniques. These in turn led to develop- 
ment of the "intelligence amplifier.” Many computers 
now approximate the human brain, both in appearance 
and in function. In I960, we started to emerge from 
childhood in the computer field, and became "teen 
agers.” Then useful standards started to appear, particu- 
larly in the field of computer to computer communica- 



COMPUTER ENGINEERS 

Engineers at the Masters and Ph.D. level with three to five 
years experience in advanced digital computer systems, 
digital techniques or control systems design are required 
to fill vacancies in the Automated Data Systems Develop- 
ment Department. 

The work will consist of system studies and advanced ap- 
paratus systems development for the automated acquisition, 
transmission, storage, processing and display of variables 
data generated in the development, design and production 
of atomic ordnance. The systems and applications being 
relatively new in concept offer unlimited horizons for highly 
creative and imaginative individuals. 

Sandia Corporation, located in Albuquerque, N.M., is en- 
gaged in research and development of nuclear weapons and 
other projects for the AEC. Albuquerque is a modern city 
of about 225,000; has an excellent climate and many cul- 
tural and recreational attractions. Winters are mild, sum- 
mer nights are cool, and there's plenty of year-around sun- 
shine. Sandia’s liberal employee benefits include generous 
vacations, retirement and insurance plans, and a graduate 
education assistance program. Paid relocation allowance. 
Send resume to Staff Employment Section 520. 

S/MM D IA 

■n* CORPORATION 



ALBUQUERQUE. NEW MEXICO 



tions. Throughout most of the thirty-five year period 
noted above, emphasis was too heavily directed toward 
preferred solutions to problems, and too little effort was 
spent in defining the problems. "Crash” proposal writ- 
ing and the professional proposal writer became out- 
moded. Complex systems, taking years to understand 
are now planned by specialists who have served the 
proper apprenticeship, know the "pitfalls,” and how to 
avoid them. 

Twenty years ago, computers were proudly made larg- 
er and faster, without proper consideration of the ability 
of the human operators to make use of their speed and 
size. Many functions were planned for computer action 
which should have been left out altogether. Some large- 
scale programs, using computers, were eventually phased 
out because the bandwidth-time product economics of 
the requisite communication nets rendered them econ- 
omically unsound. 

System Philosophy 

Early emphasis on component and subsystem reliabili- 
ty programs led to the growing realization that the sys- 
tem philosophy itself had to be sound. The modus 
operandi of the traffic engineer in the common carrier 
communication field was adopted by the value engineer 
(the V-E) in the netted computer field. This led to the 
adoption of signal and clocking standards, whereby 
computers in many locations performing many special- 
ized tasks could talk to each other over common carrier 
networks. Computer reliability and economics were bet- 
ter served through the deletion of redundant functions 
and the elimination of the need for routine inputs to 
storage by human operators. 

Present Air Traffic Control System 

All of the above contributed to the present Air Traffic 
Control System, and to those aspects of it which I shall 
now describe. 

In the overall system, a large number of computers 
now interchange data in coded form over a common 
carrier ground communication point-to-point network. 
This network is tied to the aircraft’s central computer 
system through automatic selectively addressed ground- 
to-air and air-to-ground radio communication circuits. 
Many important functions are performed to the end that 
safety, expeditiousness, comfort, and economy of flight 
are vastly improved. In line with these improvements, 
the variables and contingencies of flight are brought 
under control. The repetitive human burden is reduced 
in relation to dispatching, flight following, meteorologi- 
cal forecasting, air traffic control, and pilot position re- 
porting. 

Back in the late 1950’s, some airlines were using digi- 
tal computers to determine the best weather and flight 
conditions for long haul flights, meteorological groups 
were starting to implement plans for regional and local 
forecasting, and for automatically distributing weather 
data over large continental areas. Vertical sensing of 
meteorological data predominated, and meteorological 
observations through the use of satellites were a dream. 
No real programs for the use of aircraft as automatic 
weather sensing stations had yet been implemented. In 
air traffic control, the Federal Aviation Agency, working 
with the General Precision Laboratory, Librascope, In- 
corporated, and Tasker Instruments Corporation, was 

COMPUTERS and AUTOMATION for October, 1959 






Answer: All but this One. 



Here is a medium- scale general purpose digital computer, and 
the only small-scale thing about it is the price. It’s the low-price 
Bendix G-15. 

Part of the remarkable success of the G-15 is due to that low 
price. One user writes, . . the nearest competitor to the G-15 
costs five times as much” 

There are many other reasons why hundreds of computer experts 
have chosen the G-15. Versatility is one. The G-15 is well-equipped 
to handle all phases of business data processing as well as highly 
complex scientific and engineering calculations. Also, simplified pro- 
gramming methods allow your present personnel to solve their own 
problems on the computer. 

A broad line of accessories means the G-15 is expandable. As 
the work increases, you can add magnetic tape units, punched 
card equipment, and other accessories. Don’t forget, however, 
that the basic G-15, with its photoelectric paper tape reader- 
punch and electric typewriter, is more than adequate for most 
problems. 

Other features you will like are: true alphanumeric input-output, 
high internal speed, buffering, large program library and user’s 
exchange group, and fast nation-wide service. 



Complete details will be sent upon request. Mention specific problems 
where the G-15 may be helpful, and we will be glad to advise you. 

iptifer D,v,s,ON » dept, c-14, los angeles 45, California 

£££? | j Experienced programmers, sales and service personnel . . . 

share our growth. Contact O. P. Stdderman, Director of Marketing. 



COMPUTERS and AUTOMATION for October, 1959 




making the first really serious effort to use computers 
to improve the air traffic controller’s lot in life. First 
bold steps into automation were taking place on a num- 
ber of other aeronautical fronts as well, such as Air De- 
fense, Mission Control, Return to Base, Logistic Control, 
Aircraft Deployment, etc. Data on winds aloft, flight 
plans, and estimated times of arrival and departure, etc., 
were duplicated in a number of individual systems, and 
each group was trying to acquire data, and to send it to 
the others. But no comprehensive planning structure 
existed, no systems standards had been drafted, and 
each specialist group operated with a high degree of 
autonomy. About ten years later, in the late 1960’s, 
each group realized that independent action led to over- 
design, excessive cost, and below-par performance. It 
was then that plans were drafted for the comprehensive 
computer system which we now have. Each segment 
now confines major emphasis to its own field of interest 
and area of specialization, and draws heavily from other 
storage systems which contain data which he needs. 
Each sends high quality data to the other requiring or 
desiring it; each now makes the other’s job easier. 

Standard Air-Borne Digital Computer 

Based upon an FAA development program, there is 
now a standard air-borne low cost minimal central digital 
computer available. It is required in all cross country air- 
craft. The low cost results from the fact that the design 
was accomplished at FAA cost and production is done 
in a computer-controlled automatic factory. The air- 
craft’s air data, air navigation, and anti-collision sensors 
are used as airborne inputs. The selectively addressed 
coded communication system in the aircraft relays flight 
plans, weather data, air traffic control clearances, and 
mission control data automatically from the ground net. 
The central computer system corrects the heading refer- 
ence sensors to obtain a uniform 95 percent value for 
heading accuracy of less than one degree. The computer 
output system supplies inputs for the pictorial situation 
display system, as well as for the conventional navigation 
and flight instruments. The pilot is supplied with the 
information which enables him to maintain the estab- 
lished flight schedule within the limits imposed by other 
traffic and by the confliction criteria established for the 
area of flight. After the automatic position report is 
processed by the ground net, the air traffic control com- 
puter, cooperating with the airborne computer, gives 
him the best route and speed to fly, and the necessary 
navigational display to optimize the flight. 

Movement of Air Masses 

Aeronautical operations on a global basis have now 
entered the age of aero-flight, where the fullest mani- 
festations of the movement of air masses have become 
useful as a tool of flight. Air masses and jet streams are 
followed continually, and projections into future flight 
time are calculated at closely spaced intervals. Thus, the 
pilot of 1980 finds himself referenced to the motion of 
the medium in which he flies, taking maximum advan- 
tage in a manner which any sailing skipper of the 19th 
century would view with great envy. 

Through self-contained navigational sensors, of the 
Doppler, Doppler-Inertial, and Inertial types, airborne 

.39 



computers are combined to reduce the requirement for 
continuous ground/air cooperating navigational sys- 
tems. Applying the NAVATEC philosophy, they now 
serve primarily as zero-setting aids in the enroute serv- 
ice; uninterrupted service is now required of them pri- 
marily at departure and destination terminals. With 
automatic position reporting and ground-derived posi- 
tion determination integrated with the point-to-point 
and air/ground communication network and computer 
system, the pilot’s and the controller’s repetitive tasks 
in air traffic control are essentially eliminated. They 
can now devote full time and attention to decision mak- 
ing, thereby improving the safety and expeditiousness 
of flight. 

In this 1980 system, maximum use is made of a priori 
data, and communication circuits are used for the trans- 
fer of new and meaningful data only. Because of this, 
the economics are favorable to world-wide deployment, 
and the system is not limited to a particular kind and 
type of land or sea area. 

Surveillance functions have now gone 3-D. The com- 
mon surveillance radar net now provides altitude infor- 
mation, along with the normal plan position informa- 
tion. The effective power density has been increased 
through the application of maser and parametric ampli- 
fier techniques, and the application of electronic scan- 
ning, so that all aircraft regardless of size are seen at 
the indicated range limit. Further improvements have 
been made in removing ground and precipitation clut- 
ter, and the overhead cones are now so small that full 
volumetric coverage with slant range correction is 
available wherever needed. 

Air Space Reservations 

When airborne, the aircraft has its airspace reserva- 
tion cleared for it well in advance and no other aircraft 
may share it. Since the movement of the air mass is 
being carefully tracked for each regime of flight, there 
is a minimum of restraint imposed by the geographical 
characteristics of the land below, or by the characteristics 
of land-based electronic devices sited thereon. The air- 
borne craft may go to Mach 3 or more within the at- 
mosphere, leave the atmosphere and return and be 
guided to destination, or it may proceed at slow speed 
at low level without exceeding the limits of coverage or 
capability of the overall system. 

Family Air Scooters 

This is not the end of the air traffic control story — 
it is only a mid-chapter! The exotic fuels of twenty 
years ago are now superseded by equally high-perform- 
ance fuels which are easy to handle. Aircraft now ascend 
and descend at high angles ; long runways are required 
only for the obsolescent types still in use. With these 
new fuels, controllable jet thrust and turbofans have 
been combined to create all-weather cabin scooters, and 
family automobiles of the air. John Q. Citizen is now 
beginning to buy these instead of conventional wheeled 
automobiles. Detroit is near panic and will soon change 
over or lose its leadership to the West Coast. The next 
chapter in the air traffic control system engineering 
story — like the last — concerns new operating re- 
quirements, and the technical integration of the old 
with the new. 



COMPUTERS and AUTOMATION for October, 1959 




IN THE COMPUTER FIELD 

WHO? WHAT? WHERE? 

• 

Answers , 

Basic Source Information , 
Available to You from 

COMPUTERS 

and AUTOMATION 

PEOPLE: 

Mailing plates for over 18,500 com- 
puter people. We address your envelopes. 

$19.00 per M 

ORGANIZATIONS: 

The Computer Directory and Buyers’ 
Guide, 1958 (the June, 1958, issue of 
COMPUTERS and AUTOMATION) : 74o 
organization listings, 3220 product and 
service listings. $6.00 

GLOSSARY OF TERMS AND 
EXPRESSIONS: 

Over 480 careful, clear, understandable 
definitions. 4th cumulative edition. (20 
or more copies, 10% discount.) $1.00 

BACK COPIES: 

(For seven years of publication.) 

If available, $1.25 each, except Direc- 
tory issues June 1955 to 1958, $4.00 
each; June, 1958, $6.00. 

SUBSCRIPTIONS: 

U.S.A. one year $5.50; two years $10.50; 
add 50c per year for Canada, $1.00 per 
year elsewhere. 

BULK SUBSCRIPTIONS: 

These rates apply to prepaid subscrip- 
tions to COMPUTERS and AUTOMA- 
TION coming in together direct to the 
publisher. 

Bulk Subscription Rates 
{United. States) 

Rate for Each 

Number of Subscription, and 

Simultaneous Resulting Saving: 

Subscriptions One Year Two Years 

7 or more $4.20, 24% $7.25, 31% 

4 to 6 4.60, 16 8.00, 24 

3 5.00, 9 8.80, 16 

2 5.25, 5 9.55, 9 

For Canada, add 50 cents for each year; 
outside of the United States and Canada, 
add $1.00 for each year. 

Send prepaid orders or requests for 
more information to: 
COMPUTERS and AUTOMATION 
815 Washington St., Newtonville 60, Mass. 
If not satisfactory, returnable in seven days 
for full refund. 



BOOKS 

and 

OTHER 

PUBLICATIONS 

Moses M. Berlin 

Cambridge, Mass. 

W E PUBLISH HERE citations 
and brief reviews of books 
and other publications which have a 
significant relation to computers, data 
processing, and automation, and 
which have come to our attention. 
We shall be glad to report other in- 
formation in future lists if a review 
copy is sent to us. The plan of each 
entry is: author or editor / title / 
publisher or issuer / date, publication 
process, number of pages, price or 
its equivalent / comments. If you 
write to a publisher or issuer, we 
would appreciate your mentioning 
Computers and Automation. 

Gaynor, Frank / Concise Dictionary of 
Science / Philosophical Library, Inc., 
15 East 40 St., New York 16, N.Y. / 
1959, printed, 546 pp, $10.00 
This dictionary provides concise, ac- 
curate, definitions of terms and concepts 
pertaining to many branches of "science.” 
Included along with the basic sciences, 
physics, mathematics, astronomy, chemis- 
try, are such new fields as virology, en- 
symology, cytogenetics, radio-chemistry, 
solid state physics, and nucleonics. 

The author, editor of the Encyclopedia 
of Atomic Energy, includes all the stand- 
ard terms, as well as many modern or 
newly created descriptive words and sym- 
bols. 

Rosenblatt, Frank / The Perceptron: A 
Theory of Statistical Separability in 
Cognitive Systems / Cornell Aero- 
nautical Laboratory, Inc., for Office of 
Naval Research, publn PB 151247, Of- 
fice of Technical Servces, U. S. Dept, 
of Commerce, Washington 25, D.C. / 
1958, printed, 272 pp $4.00 
This book describes the author’s initial 
work on the theory of the Perceptron, a 
"discriminating machine” which repre- 
sents an advance in the design of an elec- 
tronic brain. The author’s research work 
on "Project PARA” (perceiving and rec- 
ognizing automaton) under the Navy’s 
sponsorship, is reported. 

The book discusses principles and meth- 
ods of electronic simulation of a human 
being’s brain, and biological implications. 
Rosenblatt, Frank / Two Theorems of 
Statistical Separability in the Perceptron 
/ Cornell Aeronautical Laboratory, Inc., 
for Office of Naval Research, publn PB 
151247-S, Office of Technical Services, 



COMPUTERS and AUTOMATION for October, 1959 



hi- 

ghest! 



Every NJE Solid State Transistor- 
ized Power Supply boasts the 
highest quality military or com- 
mercial components available 
(example: every transformer and 
choke is a MIL-T-27A style unit, 
every resistor is wire-wound). All 
are significantly derated under the 
most severe combination of line, 
load and ambient conditions. 
Proven reliability makes NJE your 
best Power Supply buy! 14 stand- 
ard models cover virtually every 
job. Here’s one: 




Output, volts: 0-10. 

Output, amps: 0-10. 

RMS ripple: 0.5 millivolts. 
Regulation, load: ±0.015% 
or ±1 MV. 

Regulation, line: ±0.03% 
or ±2 MV. 

Internal Impedance DC-20KC: 
0.02 ohms. 

PRICE: $590 

FOB Kenilworth, N. J. 



Write today for 
this new 16-page 
Power Supply catalog! 




CORPOR AT I O N 

20 Boright Ave., Kenilworth, N. J. 
Br 2-6000 






How big is a dropout? 



A dropout can be measured more than one way. 
Physically, the surface imperfection that causes a 
dropout is microscopic — often quite invisible to the 
naked eye. Financially, though, this molehill can 
become a mountain — may cost you thousands of 
dollars from a single error. 

That’s why our customers invariably demand per- 
fection from our EP Audiotape, the extra precision 
magnetic recording tape. They just can’t afford 
dropouts. 

Audio Devices’ battery of Automatic Certifiers 
is one of the unique means used to make sure EP 
Audiotape always meets customer specifications. 
The Automatic Certifier records and plays back 
every inch of the EP Audiotape under test. These 
tests can be so demanding that if the tape fails to 
reproduce just one test pulse out of the 40 million 
put on a single reel, the entire reel is rejected. 
There are no ifs, ands, or buts. 



This is one of many special quality-control opera- 
tions to which EP Audiotape is subjected. From raw 
material to hermetically sealed containers, every 
reel gets individual attention. 

EP Audiotape quality is so well verified by instru- 
ments like the Automatic Certifier that every reel 
is guaranteed to be defect-free ! For more informa- 
tion write for free Bulletin T112A. Write Dept. TA, 
Audio Devices, Inc., 444 Madison Avenue, New 
York 22, N. Y. 




444 Madison Ave., N. Y. 22, N. Y. 
in Hollywood: 840 N. Fairfax Ave. 

In Chicago: 5428 Milwaukee Ave. 

Export Dept.: 13 East 40th St., N. Y„ 16 
Rectifier Division: 620 E. Dyer Rd., Santa Ana, Calif. 



U. S. Dept, of Commerce, Washington 
25, D.C. / 1958, printed, 46 pp, $1.25 
This is a second report on the Percep- 
tron; it concentrates on such matters as 
probabalistic mathematics versus symbolic 
logic, the importance of perceptual pro- 
cesses for automata, the continuous trans- 
ducer neuron, and the organization of the 
Perceptron. 

The theory on which the electronic 
brain model is based — the "theory of 
statistical separability" — is elaborated 
upon. 

Perry, J. W., and A. Kent / Tools for Ma- 
chine Literature Searching / Interscience 
Publishers, Inc., 250 Fifth Ave., New 
York 1, N.Y. / 1958, printed, 992 pp, 
$27.50 

This is the first volume in a series de- 
signed to provide information on contem- 
porary achievements in library science 
and documentation. This volume is a 
guide to the understanding and effective 
use of modern automatic equipment in 
literature searching. The book covers: an 
introduction to machine literature search- 
ing systems; engineering of machine litera- 
ture searching systems ; applications and 
costs; programming for an IBM 650; pro- 
cedures for analyzing, encoding and 
searching of recorded information; and a 
thesaurus of scientific and technical terms, 
which include the "Semantic Code Dic- 
tionary," a monumental compilation of 
word-term tabulations, with accompany- 
ing codes. The dictionary comprises more 
than one third of the book, and offers a 
practical means for adapting encoding as 
the definitive method of information re- 
trieval. 

Uttley, A. M. / The Design of Conditional 
Probability Computers / Department of 
Scientific and Industrial Research, Na- 
tional Physical Laboratory, Teddington, 
Middlesex, England / 1958, photo off- 
set, 19 pp, cost? 

"Learning in an animal is seen in its 
simplest form in the conditioned reflex.” 
Computers which imitate such behaviour 
have been designed, and in this report a 
special purpose computer is described, 
which calculates conditioned probabilities. 
The computer can be extended to forecast 
the probability of future signals, and the 
past can be weighted in any desired man- 
ner. 

Using the illogical principle of induc- 
tion, the computer can specify conditional 
certainty, and compute conditional proba- 
bility. 

Dennis, J. B. / Mathematical Program- 
ming and Electrical Networks / John 
Wiley & Sons, Inc., (and The Tech- 
nology Press, Mass. Inst, of Technolo- 
gy), 440 Fourth Ave., New York 16, 
N.Y. / 1959, printed, 186 pp, $4.50 
This book offers a new approach to 
mathematical programming, based on an 
analogy with electrical networks. The 
author shows that any direct-current elec- 
trical network composed of current 
sources, voltage sources, ideal diodes and 
ideal transformers, is equivalent to a pair 
of dual linear programs. Extending this 
relation, certain consequences follow, and 
the procedures of the new approach are 
systematically explained. 



40 



COMPUTERS and AUTOMATION for October, 1959 





Allendoerter, C. B., and C. O. Oakley / 
Fundamentals of Freshman Mathematics 
/ McGraw-Hill, 330 West 42 St., New 
York 36, N.Y. / 1959, printed, 475 pp, 
$6.50 

It is the authors’ contention that a large 
part of the standard undergraduate curricu- 
lum in mathematics is obsolete. With this 
conviction in mind, and from recommen- 
dations by groups of mathematicians, the 
authors have compiled a text which rep- 
resents their ideas for a modern course of 
freshman mathematics. The text includes 
a review of Intermediate Algebra, a thor- 
ough discussion of the theory of sets, de- 
tailed treatment of simultaneous linear 
equations, vectors and matrices, and an 
intuitive approach to the concepts of 
limit and continuity. In organizing the 
text, the authors have kept in mind that 
the majority of those who take courses 
in mathematics are potential engineers and 
scientists. The material, therefore, can 
be used by a variety of groups, ranging 
from twelfth-grade high-school classes, to 
computer people who want a quick re- 
fresher on calculus and modern concepts. 
McCracken, Daniel D., Harold Weiss, and 
Tsai-Hwa Lee / Programming Business 
Computers / John Wiley & Sons, Inc., 
440 Fourth Ave., New York 16, N.Y. 
/ 1959, printed, 510 pp, $10.25 
The reader who lacks an extensive 
background in mathematics but who 
wishes to study the applications of elec- 
tronic computers to business problems, 
will find this an excellent "first book." 
The authors begin with fundamental top- 
ics, on the nature of the data-processing 
problem, the central concept of the file, 
flow charting, and the general characteris- 
tics of the computer. Subsequent chapters 
include information on coding, index 
registers, subroutines, input-output de- 
vices, programming, and economizing and 
estimating computer time. 

The first appendix discusses a mythical 
stored-program computer, DATAC, to il- 
lustrate a number of techniques described 
in the text. The second appendix includes 
a description of the binary system, and 
demonstrates number base conversion. 
The last appendix is a "Data-Processing 
Diary describing the programming of a 
file maintenance problem by people who 
had no previous experience with compu- 
ters. The book includes a glossary, bibli- 
ography, and index. 

Jeenel, Joachim / Programming for Digi- 
tal Computers / McGraw-Hill, 330 
West 42 St., New York 36, N.Y. / 
1959, printed, 517 pp, $12.00 
This book, expressly designed for read- 
ers with no previous experience in pro- 
gramming, guides the potential computer 
person to the preparation of problems 
for stored-program calculators. The ap- 
proach, a general one, consists of estab- 
lishing a framework upon which to build 
diverse programs. 

The nine chapters include information 
on stored-program coding, the languages 
employed, a number of checking proced- 
ures, and techniques in storage and input- 
output. Nine appendices follow the text, 
with such topics as number systems, mat- 
rix inversion, alternative plans for a net- 
work problem. The 63 pages of the ap- 




ocmai esgEGsa 

A rush project to imple- 
ment a high speed digital 
data handling or comput- 
ing system. 

0SHHH3ST 

Standard T-PAC digital 
modules by Computer Con- 
trol. 

Transistorized • compact 

• plug-in modules • etched 
circuits • taper pin solder- 
less connectors • one 
megacycle repetition rate 

• standard waveform 
throughout • reliable • no 
external - to - the - package 
coupling components • 
fully guaranteed • 



SCHEDULE 
TIMESAVERSs 

With T-PACs you go 
directly from your logical 
design to system wiring. 

Wiring is rapid, 
reliable, and permanent) yet, is 
easy to modify at any time. 

System debugging is routine, 
thanks to standard logical 
elements and 
standard waveforms. 



Write for product catalogs 





WESTERN DIVISION 

2251 BARRY AVENUE • LOS ANGELES 64 • CALIFORNIA 



COMPUTERS and AUTOMATION for October, 1959 



41 





giiiiiwft 



Very large-scale air-battle digital 
computer simulations are now going on 
at the Washington Research Office 
of tech/ops. Present operations 
call for top-flight mathematicians , 
mathematical statisticians, 
senior programmers, operations 
research analysts. 

These computer air battles are stochastic 
models which involve design and 
evaluation, and development of unusual 
techniques for studying sensitivity 
of these models to input changes. 
Associated activity involves design of 
advanced programming systems and 
of common language carriers which are 
expected to be independent of the 
first computer used — the computer itself 
augmenting and improving the 
language for use on later and more 
sophisticated computers. 

If challenging work, stimulating 
atmosphere, and an opportunity 
to participate in an unusual 
company/employee investment program 
interest you . . . write or wire collect: 



wanted : 

WAR 

GAME 

PLAYERS 



Kingsley S. Anders son 

Technical Operations , Incorporated 

3520 PROSPECT STREET, NORTHWEST • WASHINGTON T, D. C. 



♦Final position in the famed simultaneous exhibition at Pernau, 1910 
Nimzovich (white) vs Ryckhoff (black). 




pendices are followed by a bibliography 
and glossary. 

Sasieni, Maurice, Arthur Yaspan, and 
Lawrence Friedman / Operations Re- 
search — Methods and Problems / 
John Wiley & Sons, Inc., 440 Fourth 
Ave., New York 1 6, N.Y. / 1959, 
printed, 316 pp, price ? 

This book offers a clear treatment of 
formulating and solving mathematical 
models in operations research. It con- 
tains a great number of illustrative prob- 
lems with their solutions, and exercises, 
some with answers. The chapters cover: 
some basic topics in probability and 
sampling; inventory; replacement; wait- 
ing lines; competitive strategies; alloca- 
tion; linear programming; sequencing; 
and dynamic programming. The book 
features the mathematical approach to 
the study of operations research ; the level 
of mathematics assumed is a working 
knowledge of differential and integral 
calculus. Three appendices cover finite 
differences, differentiation of integrals, 
and "row operations” (a rapid method 
for solving simultaneous linear equa- 
tions.) 

Horsey, Eleanor F., and Laurence D. Sher- 
galis / Microminiaturization of Elec- 
tronic Assemblies / Hayden Book Co., 
Inc., 830 Third Ave., New York 22, 
N.Y. / 1959, printed, 288 pp., $11.00. 
The material in this book, most of 
which was presented at the Symposium on 
Microminiaturization of Electronic As- 
semblies, 1958, provides the first reference 
work covering new developments and 
techniques in this field. Detailed are the 
latest techniques in "two-dimensional” 
packaging and similar development. 

Six sections cover techniques in fabri- 
cating miniature working circuits, semi- 
conductors, components, circuits, missile 
systems, and micro electronics in industry. 

Zeines, Ben / Servomechanism Funda- 
mentals / McGraw-Hill Book Co., Inc., 
330 West 42 St., New York 36, N.Y. / 
1959, printed, 257 pp., $5.50. 

Written especially for students and tech- 
nicians, this clear, elementary treatment 
of servo systems is designed to explain 
the basic methods of servo systems and 
data transmission systems. Instead of ex- 
amining any specific system in detail, the 
book stresses fundamentals. 

The author, who is an instructor at 
RCA Institutes and Hofstra College, pre- 
sents an introductory chapter on control 
systems and servomechanisms, then, chap- 
ters on servo systems, synchros, servo ele- 
ments, electronic and magnetic amplifiers, 
direct-current and alternating-current ser- 
vomotors, methods for servos and meas- 
urements, and, examples of servos and ser- 
vo-systems. Twenty-two pages of appen- 
dices include magnetic amplifiers, direct- 
current motors, the theory of servomech- 
anisms, and elements of mechanical, acous- 
tic, and electrical systems. The book in- 
cludes a bibliography and index. 

Newman, Simon M. / Problems in Mech- 
anizing the Search in Examining Patent 
Applications / Office of Research and 
Development, Patent Office, U. S. De- 



COMPUTERS and AUTOMATION for October, 1959 





partment of Commerce, Washington 
25, D.C. / 1959, printed, 29 pp., 25c. 
This paper describes some of the many 
details and compleities which occur in 
the searching process for patent applica- 
tions, and outlines some of the interrelated 
problems that are being considered in re- 
search into methods for solving search- 
ing problems. Illustrative applications are 
given, describing the process for an orn- 
amental beaded necklace, a Christmas tree 
decorative ball hanger, plastic products, 
and manufactured toys. In each case the 
preliminary examination and the search is 
described. According to the author, uni- 
form coding systems will eventually be 
evolved, and the patents and publications 
in the search files will be encoded accord- 
ing to those systems. References and a 
bibliography follow the discussion. 

Current Research and Development in Sci- 
entific Documentation, no. 4 / National 
Science Foundation, Supt. of Docu- 
ments, U. S. Govt. Printing Office, 
Washington 25, D.C. / 1959, printed, 
85 pp., 15c. 

This volume contains about 75 brief 
descriptive progress reports on current re- 
search an development in scientific docu- 
mentation in various places. The descrip- 
tive statements are classified under five 
subject headings: (1) information require- 
ments and uses; (2) research on informa- 
tion storage and retrieval; (3) mechanical 
translation; (4) equipment development; 
(5) miscellaneous. The reports are com- 
piled as a service to individuals and or- 
ganizations interested in documentation, 
they include all pertinent activities in the 
U.S., as well as descriptions of foreign ac- 
tivities for which information can be 
secured. 

Introducing the Univac 1105 Data-Pro- 
cessing Computing System / Reming- 
ton Rand Univac, Div. of Sperry Rand 
Corp., 315 Fourth Ave., New York 10, 
N.Y. / 1959, printed, 55 pp., free. 
This pamphlet describes a general pur- 
pose computing system which is capable 
of efficiently handling data processing ap- 
plications, scientific applications, and syn- 
chronized automation. Among the fea- 
tures of the Univac 1105, are extremely 
high operating speeds, double buffers, 
flexible input-output facilities, large non- 
volatile internal random access storage, 
internally stored programs that are capa- 
ble of modification, and a double length 
accumulator. 

These features are described, and the 
applicability of the system to business and 
science is discussed. Two sections are in- 
cluded to discuss programming and speci- 
fications. 

Chestnut, Harold, and Robert W. Mayer 
/ Servomechanisms and Regulating 
System Design, vol. 1, second edition 
/ John Wiley & Sons, Inc., 440 Fourth 
Ave., New York 16, N.Y. / 1959, 
printed, 680 pp., $11.75. 

Keeping its objective similar to that of 
the first edition — to train design and 
application engineers in the basic princi- 
ples of feedback control — this reports 



also on recent advances in the field. Major 
additions include a chapter on the ap- 
plication of the root-locus method to the 
analysis and synthesis of control system 
design, and a chapter on the use of an 
analog computer for the solution of con- 
trol system problems. Other chapters are: 
the automatic control problem; manipula- 
tion of complex numbers; solution of lin- 
ear differential equations; LaPlace trans- 
forms for the solution of linear differen- 
tial equations; steady-state operation with 
sinusoidal driving functions; methods for 
determining system stability ; types of ser- 
vomechanism and control systems; design 
use of complex plane plot to improve sys- 
tem performance; and attenuation con- 
cepts for use in feedback control system 
design. In all, there are 16 chapters, fol- 
lowed by a bibliography, and a number 
of problems based on material in the text. 

Thurston, P. H. / Systems and Procedures 
Responsibility / Harvard Business 
School, Div. of Research, Soldiers 
Field, Boston 63, Mass. / 1959, printed, 
120 pp., $2.50. 

The author, a lecturer on Business Ad- 
ministration at the Harvard Business 
School, presents an administrative view 
of the division of responsibility between 
operating people and specialists, for work 
in systems and procedures. Chapters 
cover: essential problems of planning and 
installing a wide variety of systems and 
procedures in several manufacturing com- 
panies: conclusions reached on the 

strengths and weaknesses of specialists 
and operators; selection of the most ef- 
fective approaches to systems work; and 
general findings. 

Symbolic Logic, Boolean Algebra and the 
Design of Digital Systems / Technical 
Staff, Computer Control Company, Inc. 
/ Computer Control Co., Inc., 983 Con- 
cord St., Framingham, Mass. / 1959, 
printed, 32 pp., limited distribution. 
This publication presents some funda- 
mentals of symbolic logic as applied to 
the logical design of digital systems. After 
a brief history of the development of 
symbolic logic, number systems are dis- 
cussed — decimal and binary. 

A comparison of Boolean algebra and 
ordinary algebra is included, followed by 
certain theorems of symbolic logic. Some 
practical applications of the system of 
logic are demonstrated, and it is shown 
how a computer implements the princi- 
ples described. 

Handbook, Preferred Circuits, Navy Aero- 
nautical Electric Equipment, NAVAER 
16-1-519, Supplement no. 1 / National 
Bureau of Standards / Supt. of Docu- 
ments, Washington 25, D.C. / 1959, 
printed, 106 pp., $. 60 . 

This first supplement to "Preferred Cir- 
cuits” includes schematics of a number of 
circuits, followed by explanations of de- 
sign and uses. The circuits were derived 
after experimental work on a large num- 
ber of examples taken from both com- 
mercial and military electronic equipment. 
The supplement includes five instrument 
servo-circuits, two regulators, two high- 



COMPUTERS and AUTOMATION for October, 1959 




s | 



and at tech/ops f 
Monterey 

Research Office, in 
California, equally 
interesting and 
challenging work 
developing computer 
applications for 
planning, feasibility 
testing, and real time 
control of operational 
and logistical problems 
(for the Navy and 
industry) . . . 

Here, too, tech/ops needs 
scientists who habitually 
seek original thought 
patterns, and who 
respond to challenge. 
The over-all Company 
policies in the areas of 
fringe benefits and 
profit-sharing plans 
are unique. 

If challenge and reward 
appeal to you . . . write 
or wire collect: 

Harold E Kren 

Technical 
Operations , 
lncorporated / 




305 WEBSTER STREET 
MONTEREY, CALIFORNIA 



43 



SENIOR ENGINEERS / SCIENTISTS 




ovv 



iz 




__ •: 

2D00 a 

. . . built on a record of technological 
achievement at Sylvania’s 
Data Systems Operations 

Programmed growth, built on a solid base of technological and 
scientific accomplishment in advanced data processing areas, 
has made possible this rather spectacular expansion of 
Sylvania’s Data Systems Operations. Revolutionary commercial 
and military projects of long-term significance afford talented 
engineers and scientists the opportunity of virtually limitless 
growth with one of the industry’s most dynamic R&D organiza- 
tions, that is setting tomorrow’s standards now. 

Typical of the programs you can contribute to when you join 
Sylvania’s Data Systems Operations are: MOBIDIC — a unique 
concept in large-scale mobile, transistorized digital computers 
developed for the U.S. Army Signal Corps; DATA PROCESS- 
ING PHASE OF BMEWS— the USAF’s Ballistic Missile Early 
Warning System — plus a number of other special purpose 
computers. These senior positions are immediately available: 



SENIOR SYSTEMS ENGINEERS 

To work on digital data systems. 
Past experience (5-10 years) as 
Systems Designers or Project 
Leaders. Also openings for JR. 
SYSTEMS ENGINEERS with 
1-5 years experience. 

SENIOR MECHANICAL 
ENGINEERS 

Positions in product engineer- 
ing and design of data proc- 
essing systems. Areas: (1) 
packaging design of digital and 
analog circuits; (2) structural 
design of cabinets, racks, chassis 
and other modular packaging; 
(3) environmental control, test- 
ing and analysis; (4) design of 
consoles and displays; (5) de- 
sign of electronic test equip- 
ment; (6) systems integration. 



ENGINEER-IN-CHARGE 

BS or MS in EE; 5-10 years 
experience in general electron- 
ics, 3 of which should have been 
design experience in relation to 
data processing equipment. Will 
monitor technical progress, 
status and analysis of trouble 
areas and will assist in defining 
technical tasks. 

SENIOR ENGINEER 
BSEE plus 4-7 years diversified 
experience in radar systems, 
digital and analog circuitry, 
CRT, storage tube and related 
displays and some supervisory 
experience. Responsible for 
D&D and testing of electroni- 
cally programmed radar target 
simulators for use with large- 
scale radar data conversion 
systems. 



Please send resume to Mr. J. B. Dewing 

Data Systems Operations / SYLVANIA ELECTRONIC SYSTEMS 

A Division of 



* SYLVAN I A 




voltage supplies, a pulse a.f.c., and a sili- 
con transistor video amplifier, plus modi- 
fications of a number of circuits previous- 
ly issued. The aim of the preferred cir- 
cuits program is to reduce unnecessary 
circuit variations in military equipment. 

Proceedings of the Eastern Joint Computer 
Conference, Dec. 3-5, 1958, Philadelphia 
/ American Institute of Electrical En- 
gineers, 33 West 39 St., N.Y. / 1959, 
printed, 184 pp., $4.00. 

The theme of the conference was: Mod- 
ern Computers: Objectives, Designs, Ap- 
plications. The contents of this publica- 
tion include thirty-eight papers, not all of 
which are related to this theme. The role 
of computers in various industries and in 
the military are discussed. The design of 
computer components are covered in some 
of the papers; "New Frontiers” are dis- 
cussed in the first article by J. W. For- 
rester and in a number of other articles. 

An Experiment in Auto-Abstracting; 
Auto-Abstracts of Area 5 Conference 
Papers / Information Retrieval Research 
Dept., IBM Research Center, Yorktown 
Heights, New York, N.Y. / 1958, 
photo offset, 18 pp, cost? 
"Auto-abstracting” is a process of caus- 
ing a computer to make automatically an 
abstract of a piece of literature using for 
this purpose a frequency count of the 
words occurring, and selecting for the 
abstract those sentences which contain the 
most frequently occurring words. 

The samples of auto-abstracts contained 
in this collection are from machine-read- 
able transcripts, and are presented in the 
format furnished by an IBM 704 Com- 
puter. In appendices there are two dic- 
tionaries of words occurring in the docu- 
ments which are abstracted. Classification, 
translation, literature searching, and in- 
formation retrieval problems, comprise 
some of the topics of the abstracted pa- 
pers. 

Luhn, H. P., and Peter James / Litera- 
ture on Information Retrieval and Ma- 
chine Translation / The Service Bu- 
reau Corp., 425 Park Ave., New York 
22, N.Y. / 1958, printed, 42 pp, cost ? 
The principle of auto-indexing by 
means of "key-words-in-context” is em- 
ployed in this bibliography and index. 
The system is designed to facilitate the 
discovery of titles in a given subject area. 
It lists first the items by author. Then 
in a second part, there is an alphabetical 
listing of "key-words-in-context.” A brief 
introduction gives some of the advantages 
of the system; and a bibliography and in- 
dex are provided. 




-1 1 



COMPUTERS and AUTOMATION for October, 1959 





SURVEY 

OF 

RECENT 

ARTICLES 

Moses M. Berlin 
Cambridge, Mass. 

We publish here a survey of arti- 
cles related to computers and data 
processors, and their applications and 
implications, occurring in certain 
magazines. We seek to cover at least 
the following magazines: 

Automatic Control 
Automation 

Automation and Automatic 
Equipment News (British) 
Business Week 
Control Engineering 
Datamation 
Electronic Design 
Electronics 

Harvard Business Review 
Industrial Research 
Instruments and Control 
Systems 
ISA Journal 
Proceedings of the IRE 
Management Science 
The Office 
Scientific American 

The purpose of this type of refer- 
ence information is to help anybody 
interested in computers find articles 
of particular relation to this field in 
these magazines. 

For each article, we shall publish: 
the title of the article / the name of 
the author(s) / the magazine and 
issue where it appears / the pub- 
lisher’s name and address / two or 
three sentences telling what the 
article is about. 



Punched Tape Controls London Trains / 
R. Dell, Signal Engineer, London Trans- 
port / Control Engineering, vol. 6, no. 
8, August, 1959, p. 82 / McGraw-Hill 
Pub. Co., Inc., 330 West 42 St., New 
York 36, N.Y. 

The movements of 1300 trains a day are 
controlled by programs on punched tape. 
After destination and time of arrival of 
each train in the rapid transit system have 
been automatically checked against a re- 
corded timetable, switches and signals are 
set to guide the trains. Diagrams accom- 
pany the article, illustrating how the sys- 
tem works. 




sElfu Mm 



RCA . . . world leader in electronics ... is currently- 
expanding its electronic data processing operations 
as a result of one of the most significant breakthroughs 
in modern electronics— the all-transistor RCA 501 
system. Already the RCA 501 is being talked about 
as the world’s most efficient electronic data processing 
system; its sales curve is slanting sharply upwards. 

If you have experience in EDP sales or technical 
services, and are ready to step up to more challenging 
and rewarding assignments, investigate today the 
many new career openings at RCA. Current positions, 
dealing with medium and large-scale systems, in- 
clude the following: 

EDP SALES REPRESENTATIVE — background should 
include a thorough systems knowledge and at least 
one year of field experience with either government 
or commercial clients. 

EDP PROGRAMMERS AND METHODS ANALYSTS— 

local openings for qualified men to work closely with 
both customer and sales personnel in the develop- 
ment of specific applications, related procedures, 
and programs. 

For a strictly confidential interview with RCA man- 
agement, please send a detailed resume of your back- 
ground and personal qualifications to: 

Mr. E. C. Baggett 

Professional & Administrative Employment 
RCA, Dept. E-8J 
Bldg. 10-1 
Camden 2, N. J. 

RADIO CORPORATION of AMERICA 

ELECTRONIC DATA PROCESSING DIVISION 





COMPUTERS and AUTOMATION for October, 1959 



45 




An Ancient Greek Computer / Derek J. 
de Solla Price / Scientific American, 
vol. 200, no. 6, June ’59, p 60 / Scien- 
tific American, Inc., 415 Madison Ave., 
New York 17, N.Y. 

An ancient mechanism for astronomical 
calculation (2000 years old) in an Athens 
Museum is similar to a modern analog 
computer. It may bring further knowl- 
edge of Greek scientific technology and 
the evolution of modern science. 

Dynamic Testing of Computer Building 
Blocks / R. W. Buchanan and B. Kautz, 
Staff Res. Engineers, Denver Res. In- 
stitute, U. of Denver, Denver, Colo. / 
Electronics, vol. 32, no. 30, August 14, 
1959, p 66 / McGraw-Hill Pub. Co., 
330 West 42 St., New York 36, N.Y. 
This article discusses amega-pulse gen- 
erator which expedites dynamic testing of 
multi-input and gates, by providing a pulse 
source for testing the design of high- 
speed adders and gating circuits of various 
types. These pulse sources must be vari- 
able in frequency to remain useful, as the 
search for higher operating speeds pro- 
gresses. 

Computer Switching with Semiconductors 
and Relays / G. L. LaPorte and R. A. 
Marcotte, Product Engineers, IBM 
Corp., Essex Jet., Vt. / Electronics, vol. 
i 132, no. 33, August 14, 1959* p 64 / 
McGraw-Hill Pub. Co., Inc., 330 West 
42 St., New York 36, N. Y. 

The general considerations that influ- 
ence computer designers in their choice of 
electromechanical or electronic types of 



switches are pointed out in this article. 
Two tables list normal applications of 
electronic and electro-mechanical switches, 
and desirable computer switch characteris- 
tics and characteristics of semiconductors 
and relays. 

IRE Transactions on Electronic Computers 
/ Professional group on Electronic 
Computers, Institute of Radio Engin- 
eers, Inc., 1 East 79 St., New York 21, 
N.Y. / vol. EC-8, no. 2, June, 1959, 
printed, 260 pp, $6.4 5 (non-members) 
Published quarterly, this edition con- 
tains nineteen papers on computers and 
related topics, including four papers from 
the 195 9 Solid-State Circuits Conference, 
and six papers from the 1958 National 
Simulation Conference. Five brief arti- 
cles are included in the Correspondence 
section, and the publication furnishes ab- 
stracts of current computer literature, 
PGEC news and notices, and the Science 
Education Subcommittee Newsletter, 
SENEWS, vol. 2, no. 2, June, 1959- 

Burroughs Puts Production Control on 
205 / E. H. Goodman / Computing 
News, vol. 7, no. 15, Aug. 1, 1959, 
p 154 — 3 / C. N., P.O. Box 90424, 
Airport Station, Los Angeles 45, Calif. 
A computer manufacturer has applied 
one of its computers to production con- 
trol. The article describes the installation 
of, and some of the planned uses for, the 
computer. 

Keeping your Maintenance Records with 
a Computer / F. H. Winterkamp, In- 



strument Div. Superintendent, E. I. du- 
Pont deNemours & Co., Inc. / ISA 
Journal, vol. 6, no. 8, August, 1959, 
p 44 / Instrument Society of America, 
313 Sixth Ave., Pittsburgh 22, Pa. 
Computer analysis and control enables 
a large industrial plant to keep track of 
hundreds of instruments with diverse cal- 
ibrations, alarm settings, and preventive 
maintenance requirements. The punched- 
card records help improve preventive 
maintenance and provide lower service 
costs. 

New Pad and Pencil for Electronic Of- 
fices / Modern Office Procedures, vol. 
4, no. 8, August, 1959, p 12 / The In- 
dustrial Publishing Corp., 812 Huron 
Rd., Cleveland 15, Ohio. 

This article provides a broad view of 
the computer and what it can do. It con- 
siders magnetic tape units as the pads and 
pencils, and describes them in terms a 
layman can understand. 

Magnetic Core Circuits for Digital Appli- 
cations / I. L. Auerbach, Auerbach 
Electronics Corp. / Automatic Control, 
vol. 11, no. 2, Aug., 1959, p 48 / Rein- 
hold Pub. Corp., 430 Park Ave., New 
York 22, N.Y. 

Bistable magnetic cores, one of the 
prime component types for digital appli- 
cations, can perform most of the functions 
of digital data processing systems, includ- 
ing storage, delay, control and logical 
operations. This article discusses present 
applications of magnetic core circuits. 



ADVERTISING INDEX 



Following is the index of advertisements. Each item contains: 

Name and address of the advertiser / page number where the 

advertisement appears / name of agency if any. 

Ampex Corp., Instrumentation Div., 934 Charter St., 
Redwood City, Calif. / Page 5 / McCann Erickson, Inc. 

Audio Devices, Inc., 444 Madison Ave., New York 22, 
N.Y. / Page 40 / Marsteller, Rickard, Gebhardt & 
Reed, Inc. 

Bendix Aviation Corp., Computer Div., 5630 Arbor 
Vitae St., Los Angeles, Calif. / Page 37 / Shaw Ad- 
vertising Inc. 

Bendix Aviation Corp., Eclipse-Pioneer Div., Route 46 
at 17, Teterboro, N.J. / Page 2 / Deutsch & Shea, Inc. 

Chrysler Corp., Engineering Div., P.O. Box 1118, De- 
troit 31, Mich. / Page 31 / N. W. Ayer & Son, Inc. 

C. P. Clare & Co., 3101 Pratt Blvd., Chicago 45, 111. / 
Page 34 / Reincke, Meyer & Finn 

Computer Control Co., 983 Concord St., Framingham, 
Mass. / Page 41 / Briant Advertising 

General Electric Co., Schenectady, N.Y. / Page 47 / 
G. M. Basford Co. 

Hughes Aircraft Co., Hughes Products, Industrial Sys- 
tems Div., International Airport Station, Los Angeles 
45, Calif. / Page 7 / Foote, Cone & Belding 

Minneapolis-Honeywell Regulator Co., DATAmatic 
Div., Newton Highlands 61, Mass. / Page 8 / Batten, 
Barton, Durstine & Osborn 

The Mitre Corp., 244 Wood St., Lexington 73, Mass. / 
Page 33 / Deutsch & Shea, Inc. 

46 



NJE Corp., 345 Carnegie Ave., Kenilworth, N.J. / Page 
39 / Keyes, Martin & Co. 

Philco Corp., Government & Industrial Div., 4700 Wis- 
sahickon Ave., Philadelphia 44, Pa. / Page 3 / Max- 
well Associates, Inc. 

Radio Corp. of America, Camden, N.J. / Page 45 / A1 
Paul Lefton Co., Inc. 

Radio Corp. of America, Semiconductor and Materials 
Div., Somerville, N.J. / Page 35 / A1 Paul Lefton 
Co., Inc. 

Ramo-Wooldridge, a Div. of Thompson Ramo Wool- 
dridge Inc., P.O. Box 90534, Airport Station, Los 
Angeles 45, Calif. / Page 19 / The McCarty Co. 

Sandia Corp., Sandia Base, Albuquerque, N.M. / Page 
36 / Ward Hicks Advertising 

Space Technology Laboratories, Inc., P.O. Box 95004, 
Los Angeles 45, Calif. / Page 21 / Gaynor & 
Ducas, Inc. 

Sylvania Electronic Systems, 189 B St., Needham 94, 
Mass. / Page 44 / Deutsch & Shea, Inc. 

System Development Corp., 2406 Colorado Ave., Santa 
Monica, Calif. / Page 48 / Stromberger, LaVene, 
McKenzie 

Technical Operations, Inc., 3520 Prospect St., N.W., 
Washington 7, D.C. / Page 42 / Dawson MacLeod 
& Stivers 

Technical Operations, Inc., 305 Webster St., Monterey, 
Calif. / Page 43 / Dawson MacLeod & Stivers 

COMPUTERS and AUTOMATION for October, 1959 



Here’s how General Electric solves 
typical DC power-supply problems 

for computers and special applications 



PROBLEM 



SOLUTION 



“We need to devote our engineering time 
to designing our electronic circuitry . . . 
not the power components.” 




This is a frequent problem facing computer manu- 
facturers. General Electric’s Rectifier Department has 
complete engineering and manufacturing capability 
not only to design and apply all types of power sup- 
plies, but also to incorporate power supplies into 
completely integrated systems. 

These systems could include load distribution, sup- 
ply sequencing, protection for power supply and load, 
and complete power distribution. Let General Electric 
tackle your DC power problems such as those asso- 
ciated with load IR drop, “cross talk,” and other 
nuisance-type problems plaguing your engineers. 



PROBLEM 



“it’s always a problem making 
sure transistorized equipment is 
safe from its power supply.” 



SOLUTION 



PROBLEM 



“ My power supply requirements 
fluctuate so much . . . big jobs, 
little jobs, all in between 



SOLUTION 



PROBLEM 



“We have a real low-voltage 
power distribution problem with 
our computer. ” 



SOLUTION 



To alleviate this problem, General 
Electric has developed several meth- 
ods of making transistorized equip- 
ment safer in this respect. With G-E 
protective circuits, shorting a plus 
high-voltage bus to a plus or minus 
low-voltage bus would not cause the 
low-voltage bus to exceed a small 
percentage of nominal rated value. 

General Electric power supplies 
protect completely transistorized 
pieces of equipment from large losses 
due to over-voltage failures. 



G.E. has built individual power sup- 
plies and complete systems ranging 
from less than one watt up to 35,000 
kilowatts. These power supplies span 
the complete range of DC power — 
regulated and unregulated — applying 
all types of components. G-E expe- 
rience includes completely transistor- 
ized supplies, and supplies with the 
new controlled rectifier, magnetic 
amplifiers, voltage stabilizing trans- 
formers, and motor-alternator “brute 
force” systems. 




Low -voltage dis- 
tribution prob- 
lems can be han- 
dledeasily through 
load compensa- 
tion. Curve “A” 
is net desired no- 
load to full-load regulation at load 
point “B” is regulation at load with- 
out remote sensing or load compensa- 
tion. “C” represents IR compensation 
in power supply itself. “D” is amount 
of IR or load compensation. 



NO MATTER WHAT your computer and other 
special power-supply problems are, General 
Electric can help you economize — economize 
by helping you free your engineers of these 
problems. For more information on power- 



supply products and services, contact your 
nearest General Electric Apparatus Sales 
Office or write to Section G 535-2, General 
Electric Company, Schenectady, New York. 



Ptogress Is Our Most Important "Product 



GENERAL 




ELECTRIC 



COMPUTERS and AUTOMATION for October, 1959 



47 






“COMPUTER PROGRAMMING at SDC is a fundamental discipline rather than a service. This 
approach to programming reflects the special nature of SDC's work-developing large-scale computer-centered systems. 
"Our computing facility is the largest in the world. Our work includes programming for real time systems, studies of 
automatic programming, machine translation, pattern recognition , information retrieval, simulation, and a variety of other 
data processing problems. SDC is one of the few organizations that carries on such broad research and development 
in programming. 

" When we consider a complex system that involves a high speed computer, we look on the computer program as a 
system component— one requiring the same attention as the hardware, and designed to mesh with other components. 

We feel that the program must not simply be patched in later. This point of view means that SDC programmers are 
participants in the development of a system and that they influence the design of components such as computers and 
communication links, in much the same way as hardware design influences computer programs. 



"Major expansion in our work has created a number of new positions for those who wish to accept new challenges 
in programming. Senior positions are open. / suggest you write directly to Mr. William Keefer at the address below. 
He is responsible for prompt response io your correspondence.' ’ 



e>. s>w& 




Senior Computer Systems Specialist